Image classification is a popular machine learning based applications of deep learning. Deep learning techniques are very popular because they can be effectively used in performing operations on image data in large-scale. In this paper CNN model was designed to better classify images. We make use of feature extraction part of inception v3 model for feature vector calculation and retrained the classification layer with these feature vector. By using the transfer learning mechanism the classification layer of the CNN model was trained with 20 classes of Caltech101 image dataset and 17 classes of Oxford 17 flower image dataset. After training, network was evaluated with testing dataset images from Oxford 17 flower dataset and Caltech101 image dataset. The mean testing precision of the neural network architecture with Caltech101 dataset was 98 % and with Oxford 17 Flower image dataset was 92.27 %.
A Novel GA-SVM Model For Vehicles And Pedestrial Classification In Videosijtsrd
The paper presents a novel algorithm for object classification in videos based on improved support vector machine (SVM) and genetic algorithm. One of the problems of support vector machine is selection of the appropriate parameters for the kernel. This has affected the accuracy of the SVM over the years. This research aims at optimizing the SVM Radial Basis kernel parameters using the genetic algorithm. Moving object classification is a requirement in smart visual surveillance systems as it allows the system to know the kind of object in the scene and be able to recognize the actions the object can perform. This paper presents an GA-SVM machine learning approach for real time object classification in videos. Radial distance signal features are extracted from the silhouettes of object detected in videos. The radial distance signals features are then normalized and fed into the GA-SVM model. The classification rate of 99.39% is achieved with the genetically trained SVM algorithm while 99.1% classification accuracy is achieved with the normal SVM. A comparison of this classifier with some other classifiers in terms of classification accuracy shows a better performance than other classifiers such as the normal SVM, Artificial neural network (ANN), Genetic Artificial neural network (GANN), K-nearest neighbor (K-NN) and K-Means classifiers. Akintola Kolawole G."A Novel GA-SVM Model For Vehicles And Pedestrial Classification In Videos" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-1 | Issue-4 , June 2017, URL: http://www.ijtsrd.com/papers/ijtsrd109.pdf http://www.ijtsrd.com/computer-science/artificial-intelligence/109/a-novel-ga-svm-model-for-vehicles-and-pedestrial-classification-in-videos/akintola-kolawole-g
Comparison Between Levenberg-Marquardt And Scaled Conjugate Gradient Training...CSCJournals
The document compares the Levenberg-Marquardt and Scaled Conjugate Gradient algorithms for training a multilayer perceptron neural network for image compression. It finds that while both algorithms performed comparably in terms of accuracy and speed, the Levenberg-Marquardt algorithm achieved slightly better accuracy as measured by average training accuracy and mean squared error, while the Scaled Conjugate Gradient algorithm was faster as measured by average training iterations. The document compresses a standard test image called Lena using both algorithms and analyzes the results.
Image fusion is a sub field of image processing in which more than one images are fused to create an image where all the objects are in focus. The process of image fusion is performed for multi-sensor and multi-focus images of the same scene. Multi-sensor images of the same scene are captured by different sensors whereas multi-focus images are captured by the same sensor. In multi-focus images, the objects in the scene which are closer to the camera are in focus and the farther objects get blurred. Contrary to it, when the farther objects are focused then closer objects get blurred in the image. To achieve an image where all the objects are in focus, the process of images fusion is performed either in spatial domain or in transformed domain. In recent times, the applications of image processing have grown immensely. Usually due to limited depth of field of optical lenses especially with greater focal length, it becomes impossible to obtain an image where all the objects are in focus. Thus, it plays an important role to perform other tasks of image processing such as image segmentation, edge detection, stereo matching and image enhancement. Hence, a novel feature-level multi-focus image fusion technique has been proposed which fuses multi-focus images. Thus, the results of extensive experimentation performed to highlight the efficiency and utility of the proposed technique is presented. The proposed work further explores comparison between fuzzy based image fusion and neuro fuzzy fusion technique along with quality evaluation indices.
Comparative Performance of Image Scrambling in Transform Domain using Sinusoi...CSCJournals
With the rapid development of technology, and the popularization of internet, communication is been greatly promoted. The communication is not limited only to information but also includes multimedia information like digital Images. Therefore, the security of digital images has become a very important and practical issue, and appropriate security technology is used for those digital images containing confidential or private information especially. In this paper a novel approach of Image scrambling has been proposed which includes both spatial as well as Transform domain. Experimental results prove that correlation obtained in scrambled images is much lesser then the one obtained in transformed images.
This document summarizes a research paper that compares different classification techniques for remotely sensed Landsat images. It discusses extracting textural features using GLCM, spatial features using PSI and SFS, and applying feature selection using S-Index. Supervised neural network classifiers like k-NN, BPNN and PCNN are tested on a Landsat image of Brazil and compared to unsupervised techniques. Results show BPNN achieved the highest classification accuracy.
A NOVEL IMAGE STEGANOGRAPHY APPROACH USING MULTI-LAYERS DCT FEATURES BASED ON...ijma
Steganography is the science of hidden data in the cover image without any updating of the cover image.
The recent research of the steganography is significantly used to hide large amount of information within
an image and/or audio files. This paper proposed a new novel approach for hiding the data of secret image
using Discrete Cosine Transform (DCT) features based on linear Support Vector Machine (SVM)
classifier. The DCT features are used to decrease the image redundant information. Moreover, DCT is
used to embed the secrete message based on the least significant bits of the RGB. Each bit in the cover
image is changed only to the extent that is not seen by the eyes of human. The SVM used as a classifier to
speed up the hiding process via the DCT features. The proposed method is implemented and the results
show significant improvements. In addition, the performance analysis is calculated based on the
parameters MSE, PSNR, NC, processing time, capacity, and robustness.
IRJET- Mango Classification using Convolutional Neural NetworksIRJET Journal
This document presents research on classifying different types of mangoes using convolutional neural networks (CNNs). The researchers collected a dataset of over 5000 mango images across 5 classes. They used transfer learning with the Inception v3 CNN model pre-trained on ImageNet, removing the final classification layer and retraining a new one for the mango classes. The CNN achieved over 99% accuracy on the test set at classifying mango types, demonstrating that CNNs can effectively perform fine-grained image classification of mangoes and distinguish between similar types.
This document proposes a convolutional neural network (CNN) to automatically classify aerial and remote sensing images. The CNN has six layers - three convolutional layers to extract visual features from the images at different levels of abstraction, two fully-connected layers to integrate the extracted features, and a final softmax classifier layer to classify the images. The CNN is evaluated on two datasets and is shown to outperform state-of-the-art baselines in terms of classification accuracy, demonstrating its ability to learn spatial features directly from images without relying on handcrafted features or descriptors.
A Novel GA-SVM Model For Vehicles And Pedestrial Classification In Videosijtsrd
The paper presents a novel algorithm for object classification in videos based on improved support vector machine (SVM) and genetic algorithm. One of the problems of support vector machine is selection of the appropriate parameters for the kernel. This has affected the accuracy of the SVM over the years. This research aims at optimizing the SVM Radial Basis kernel parameters using the genetic algorithm. Moving object classification is a requirement in smart visual surveillance systems as it allows the system to know the kind of object in the scene and be able to recognize the actions the object can perform. This paper presents an GA-SVM machine learning approach for real time object classification in videos. Radial distance signal features are extracted from the silhouettes of object detected in videos. The radial distance signals features are then normalized and fed into the GA-SVM model. The classification rate of 99.39% is achieved with the genetically trained SVM algorithm while 99.1% classification accuracy is achieved with the normal SVM. A comparison of this classifier with some other classifiers in terms of classification accuracy shows a better performance than other classifiers such as the normal SVM, Artificial neural network (ANN), Genetic Artificial neural network (GANN), K-nearest neighbor (K-NN) and K-Means classifiers. Akintola Kolawole G."A Novel GA-SVM Model For Vehicles And Pedestrial Classification In Videos" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-1 | Issue-4 , June 2017, URL: http://www.ijtsrd.com/papers/ijtsrd109.pdf http://www.ijtsrd.com/computer-science/artificial-intelligence/109/a-novel-ga-svm-model-for-vehicles-and-pedestrial-classification-in-videos/akintola-kolawole-g
Comparison Between Levenberg-Marquardt And Scaled Conjugate Gradient Training...CSCJournals
The document compares the Levenberg-Marquardt and Scaled Conjugate Gradient algorithms for training a multilayer perceptron neural network for image compression. It finds that while both algorithms performed comparably in terms of accuracy and speed, the Levenberg-Marquardt algorithm achieved slightly better accuracy as measured by average training accuracy and mean squared error, while the Scaled Conjugate Gradient algorithm was faster as measured by average training iterations. The document compresses a standard test image called Lena using both algorithms and analyzes the results.
Image fusion is a sub field of image processing in which more than one images are fused to create an image where all the objects are in focus. The process of image fusion is performed for multi-sensor and multi-focus images of the same scene. Multi-sensor images of the same scene are captured by different sensors whereas multi-focus images are captured by the same sensor. In multi-focus images, the objects in the scene which are closer to the camera are in focus and the farther objects get blurred. Contrary to it, when the farther objects are focused then closer objects get blurred in the image. To achieve an image where all the objects are in focus, the process of images fusion is performed either in spatial domain or in transformed domain. In recent times, the applications of image processing have grown immensely. Usually due to limited depth of field of optical lenses especially with greater focal length, it becomes impossible to obtain an image where all the objects are in focus. Thus, it plays an important role to perform other tasks of image processing such as image segmentation, edge detection, stereo matching and image enhancement. Hence, a novel feature-level multi-focus image fusion technique has been proposed which fuses multi-focus images. Thus, the results of extensive experimentation performed to highlight the efficiency and utility of the proposed technique is presented. The proposed work further explores comparison between fuzzy based image fusion and neuro fuzzy fusion technique along with quality evaluation indices.
Comparative Performance of Image Scrambling in Transform Domain using Sinusoi...CSCJournals
With the rapid development of technology, and the popularization of internet, communication is been greatly promoted. The communication is not limited only to information but also includes multimedia information like digital Images. Therefore, the security of digital images has become a very important and practical issue, and appropriate security technology is used for those digital images containing confidential or private information especially. In this paper a novel approach of Image scrambling has been proposed which includes both spatial as well as Transform domain. Experimental results prove that correlation obtained in scrambled images is much lesser then the one obtained in transformed images.
This document summarizes a research paper that compares different classification techniques for remotely sensed Landsat images. It discusses extracting textural features using GLCM, spatial features using PSI and SFS, and applying feature selection using S-Index. Supervised neural network classifiers like k-NN, BPNN and PCNN are tested on a Landsat image of Brazil and compared to unsupervised techniques. Results show BPNN achieved the highest classification accuracy.
A NOVEL IMAGE STEGANOGRAPHY APPROACH USING MULTI-LAYERS DCT FEATURES BASED ON...ijma
Steganography is the science of hidden data in the cover image without any updating of the cover image.
The recent research of the steganography is significantly used to hide large amount of information within
an image and/or audio files. This paper proposed a new novel approach for hiding the data of secret image
using Discrete Cosine Transform (DCT) features based on linear Support Vector Machine (SVM)
classifier. The DCT features are used to decrease the image redundant information. Moreover, DCT is
used to embed the secrete message based on the least significant bits of the RGB. Each bit in the cover
image is changed only to the extent that is not seen by the eyes of human. The SVM used as a classifier to
speed up the hiding process via the DCT features. The proposed method is implemented and the results
show significant improvements. In addition, the performance analysis is calculated based on the
parameters MSE, PSNR, NC, processing time, capacity, and robustness.
IRJET- Mango Classification using Convolutional Neural NetworksIRJET Journal
This document presents research on classifying different types of mangoes using convolutional neural networks (CNNs). The researchers collected a dataset of over 5000 mango images across 5 classes. They used transfer learning with the Inception v3 CNN model pre-trained on ImageNet, removing the final classification layer and retraining a new one for the mango classes. The CNN achieved over 99% accuracy on the test set at classifying mango types, demonstrating that CNNs can effectively perform fine-grained image classification of mangoes and distinguish between similar types.
This document proposes a convolutional neural network (CNN) to automatically classify aerial and remote sensing images. The CNN has six layers - three convolutional layers to extract visual features from the images at different levels of abstraction, two fully-connected layers to integrate the extracted features, and a final softmax classifier layer to classify the images. The CNN is evaluated on two datasets and is shown to outperform state-of-the-art baselines in terms of classification accuracy, demonstrating its ability to learn spatial features directly from images without relying on handcrafted features or descriptors.
IRJET- Deep Convolutional Neural Network for Natural Image Matting using Init...IRJET Journal
This document describes a study that used a deep convolutional neural network (CNN) to perform natural image matting using initial alpha mattes. The researchers trained a CNN using alpha mattes generated from closed form matting and K-nearest neighbor (KNN) matting as inputs. Combining the results from these two existing matting methods, which take different local image structures as input, achieved more accurate foreground extraction than prior methods alone. The CNN was able to classify images with higher performance than existing algorithms by using convolutional layers instead of fully connected layers.
This document summarizes a research paper about developing a new set of low-complexity features for detecting steganography in JPEG images. The proposed features, called DCTR features, are computed by taking the discrete cosine transform (DCT) of non-overlapping 8x8 blocks of the image, resulting in 64 feature maps. Histograms are formed from the quantized noise residuals in these feature maps. This approach has lower computational complexity than previous rich models used for steganalysis and provides competitive detection accuracy across different steganographic algorithms while using fewer features. The paper introduces the concept of an undecimated DCT and explains how it relates to previous work in JPEG steganalysis.
Optimized Neural Network for Classification of Multispectral ImagesIDES Editor
This document summarizes an article that proposes using a multiobjective particle swarm optimization (MOPSO) approach to optimize the structure of an artificial neural network for classifying multispectral satellite images. Specifically, the MOPSO is used to simultaneously select the most discriminative spectral bands from the available options and determine the optimal number of nodes in the hidden layer of the neural network. The MOPSO approach is compared to traditional classifiers like maximum likelihood classification and Euclidean classifiers. The results show that the MOPSO-optimized neural network approach provides superior performance for remote sensing image classification problems.
High Speed Data Exchange Algorithm in Telemedicine with Wavelet based on 4D M...Dr. Amarjeet Singh
Existing Medical imaging techniques such as fMRI, positron emission tomography (PET), dynamic 3D ultrasound and dynamic computerized tomography yield large amounts of four-dimensional sets. 4D medical data sets are the series of volumetric images netted in time, large in size and demand a great of assets for storage and transmission. Here, in this paper, we present a method wherein 3D image is taken and Discrete Wavelet Transform(DWT) and Dual-Tree Complex Wavelet Transform(DTCWT) techniques are applied separately on it and the image is split into sub-bands. The encoding and decoding are done using 3D-SPIHT, at different bit per pixels(bpp). The reconstructed image is synthesized using Inverse DWT technique. The quality of the compressed image has been evaluated using some factors such as Mean Square Error(MSE) and Peak-Signal to Noise Ratio (PSNR).
IRJET- Face Recognition using Machine LearningIRJET Journal
This document presents a modified CNN architecture for face recognition that adds two batch normalization operations to improve performance. The CNN extracts facial features using convolutional layers and max pooling, and classifies faces using a softmax classifier. The proposed approach was tested on a face database containing images of 4 individuals with varying lighting conditions. Experimental results showed the modified CNN with batch normalization achieved better recognition results than traditional methods.
A new block cipher for image encryption based on multi chaotic systemsTELKOMNIKA JOURNAL
In this paper, a new algorithm for image encryption is proposed based on three chaotic systems which are Chen system,logistic map and two-dimensional (2D) Arnold cat map. First, a permutation scheme is applied to the image, and then shuffled image is partitioned into blocks of pixels. For each block, Chen system is employed for confusion and then logistic map is employed for generating subsititution-box (S-box) to substitute image blocks. The S-box is dynamic, where it is shuffled for each image block using permutation operation. Then, 2D Arnold cat map is used for providing diffusion, after that XORing the result using Chen system to obtain the encrypted image.The high security of proposed algorithm is experimented using histograms, unified average changing intensity (UACI), number of pixels change rate (NPCR), entropy, correlation and keyspace analyses.
Image Steganography Using Wavelet Transform And Genetic AlgorithmAM Publications
This paper presents the application of Wavelet Transform and Genetic Algorithm in a novel
steganography scheme. We employ a genetic algorithm based mapping function to embed data in Discrete Wavelet
Transform coefficients in 4x4 blocks on the cover image. The optimal pixel adjustment process is applied after
embedding the message. We utilize the frequency domain to improve the robustness of steganography and, we
implement Genetic Algorithm and Optimal Pixel Adjustment Process to obtain an optimal mapping function to
reduce the difference error between the cover and the stego-image, therefore improving the hiding capacity with
low distortions. Our Simulation results reveal that the novel scheme outperforms adaptive steganography technique
based on wavelet transform in terms of peak signal to noise ratio and capacity, 39.94 dB and 50% respectively.
IRJET- Identification of Scene Images using Convolutional Neural Networks - A...IRJET Journal
This document summarizes research on using convolutional neural networks (CNNs) for scene image identification. It first discusses traditional object detection methods and their limitations. CNNs are presented as an improved approach, with convolutional, pooling and fully connected layers to extract features and classify images. Several popular CNN-based object detection algorithms are then summarized, including R-CNN, Fast R-CNN, Faster R-CNN and YOLO. The document concludes that CNN methods provide more accurate object identification compared to traditional algorithms due to their ability to learn from large datasets.
This document is a project report submitted by Shubham Jain and Vikas Jain for their course CS676A. The project aims to learn relative attributes associated with face images using the PubFig dataset. Convolutional neural network features and the RankNet model were used to predict attribute rankings. RankNet achieved better performance than RankSVM and GIST features. Zero-shot learning for unseen classes was explored by building probabilistic class models, but performance was poor. Future work could improve the modeling of unseen classes.
Architecture neural network deep optimizing based on self organizing feature ...journalBEEI
Forward neural network (FNN) execution relying on the algorithm of training and architecture selection. Different parameters using for nip out the architecture of FNN such as the connections number among strata, neurons hidden number in each strata hidden and hidden strata number. Feature architectural combinations exponential could be uncontrollable manually so specific architecture can be design automatically by using special algorithm which build system with ability generalization better. Determination of architecture FNN can be done by using the algorithm of optimization numerous. In this paper methodology new proposes achievement where FNN neurons respective with hidden layers estimation work where in this work collect algorithm training self organizing feature map (SOFM) with advantages to explain how the best architectural selected automatically by SOFM from criteria error testing based on architecture populated. Different size of dataset benchmark of 4 classifications tested for approach proposed.
This document summarizes a study that used neural networks and particle swarm optimization incorporating fuzzy c-means (PSOFCN) segmentation to recognize handwritten characters in the Meetei Mayek script. 34 characters were analyzed. Images were preprocessed, segmented using PSOFCN and recognized using a multilayer feedforward neural network with backpropagation. 1700 samples were used for training and 1700 for testing. Recognition accuracy ranged from 30-100%, with an average of 72%. Characters with simpler shapes had higher accuracy than more complex characters.
International Journal of Research in Engineering and Science is an open access peer-reviewed international forum for scientists involved in research to publish quality and refereed papers. Papers reporting original research or experimentally proved review work are welcome. Papers for publication are selected through peer review to ensure originality, relevance, and readability.
Particle Swarm Optimization for Nano-Particles Extraction from Supporting Mat...CSCJournals
This document summarizes a study that uses Particle Swarm Optimization (PSO) for automatic segmentation of nano-particles in Transmission Electron Microscopy (TEM) images. PSO is applied to specify local and global thresholds for segmentation by treating image entropy as a minimization problem. Results show the PSO method improves over previous techniques by reducing incorrect characterization of nano-particles in images affected by liquid concentrations or supporting materials, with up to a 27% reduction in errors. Compared to manual characterization, PSO provides comparable particle counting with higher computational efficiency suitable for real-time analysis.
The document discusses clustering images based on their properties. Images are converted into intensity, contrast, Weibull and fractal images. Eight properties are calculated for each image type, including brightness, standard deviation, entropy, skewness, kurtosis, separability, spatial frequency and visibility. The properties are normalized and clustered using k-means clustering. Tables show normalized property values for different image types. The clustering groups similar images based on their discriminative properties.
Multilayer extreme learning machine for hand movement prediction based on ele...journalBEEI
Brain computer interface (BCI) technology connects humans with machines via electroencephalography (EEG). The mechanism of BCI is pattern recognition, which proceeds by feature extraction and classification. Various feature extraction and classification methods can differentiate human motor movements, especially those of the hand. Combinations of these methods can greatly improve the accuracy of the results. This article explores the performances of nine feature-extraction types computed by a multilayer extreme learning machine (ML-ELM). The proposed method was tested on different numbers of EEG channels and different ML-ELM structures. Moreover, the performance of ML-ELM was compared with those of ELM, Support Vector Machine and Naive Bayes in classifying real and imaginary hand movements in offline mode. The ML-ELM with discrete wavelet transform (DWT) as feature extraction outperformed the other classification methods with highest accuracy 0.98. So, the authors also found that the structures influenced the accuracy of ML-ELM for different task, feature extraction used and channel used.
This document presents a method for image upscaling using a fuzzy ARTMAP neural network. It begins with an introduction to image upscaling and interpolation techniques. It then provides background on ARTMAP neural networks and fuzzy logic. The proposed method uses a linear interpolation algorithm trained with an ARTMAP network. Results show the method performs better than nearest neighbor interpolation in terms of peak signal-to-noise ratio, mean squared error, and structural similarity, though not as high as bicubic interpolation. Overall, the fuzzy ARTMAP network provides an effective way to perform image upscaling with fewer artifacts than traditional methods.
Semantic Concept Detection in Video Using Hybrid Model of CNN and SVM Classif...CSCJournals
In today's era of digitization and fast internet, many video are uploaded on websites, a mechanism is required to access this video accurately and efficiently. Semantic concept detection achieve this task accurately and is used in many application like multimedia annotation, video summarization, annotation, indexing and retrieval. Video retrieval based on semantic concept is efficient and challenging research area. Semantic concept detection bridges the semantic gap between low level extraction of features from key-frame or shot of video and high level interpretation of the same as semantics. Semantic Concept detection automatically assigns labels to video from predefined vocabulary. This task is considered as supervised machine learning problem. Support vector machine (SVM) emerged as default classifier choice for this task. But recently Deep Convolutional Neural Network (CNN) has shown exceptional performance in this area. CNN requires large dataset for training. In this paper, we present framework for semantic concept detection using hybrid model of SVM and CNN. Global features like color moment, HSV histogram, wavelet transform, grey level co-occurrence matrix and edge orientation histogram are selected as low level features extracted from annotated groundtruth video dataset of TRECVID. In second pipeline, deep features are extracted using pretrained CNN. Dataset is partitioned in three segments to deal with data imbalance issue. Two classifiers are separately trained on all segments and fusion of scores is performed to detect the concepts in test dataset. The system performance is evaluated using Mean Average Precision for multi-label dataset. The performance of the proposed framework using hybrid model of SVM and CNN is comparable to existing approaches.
GeoAI: A Model-Agnostic Meta-Ensemble Zero-Shot Learning Method for Hyperspec...Konstantinos Demertzis
The document discusses a new meta-ensemble zero-shot learning method called MAME-ZsL for hyperspectral image analysis and classification. MAME-ZsL overcomes the difficulties of traditional deep learning methods that require large labeled datasets and long training times. It reduces computational costs, avoids overfitting, and achieves high classification accuracy even when testing classes were not present during training. The method is a novel optimization-based meta-ensemble architecture that facilitates learning representations from limited labeled examples to enable one-shot and zero-shot learning.
We presents a technique for moving objects extraction. There are several different approaches for moving object extraction, clustering is one of object extraction method with a stronger teorical foundation used in many applications. And need high performance in many extraction process of moving object. We compare K-Means and Self-Organizing Map method for extraction moving objects, for performance measurement of moving object extraction by applying MSE and PSNR. According to experimental result that the MSE value of K-Means is smaller than Self-Organizing Map. It is also that PSNR of K-Means is higher than Self-Organizing Map algorithm. The result proves that K-Means is a promising method to cluster pixels in moving objects extraction.
CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION cscpconf
This paper aims at providing insight on the transferability of deep CNN features to
unsupervised problems. We study the impact of different pretrained CNN feature extractors on
the problem of image set clustering for object classification as well as fine-grained
classification. We propose a rather straightforward pipeline combining deep-feature extraction
using a CNN pretrained on ImageNet and a classic clustering algorithm to classify sets of
images. This approach is compared to state-of-the-art algorithms in image-clustering and
provides better results. These results strengthen the belief that supervised training of deep CNN
on large datasets, with a large variability of classes, extracts better features than most carefully
designed engineering approaches, even for unsupervised tasks. We also validate our approach
on a robotic application, consisting in sorting and storing objects smartly based on clustering
Plant Disease Detection using Convolution Neural Network (CNN)IRJET Journal
This document describes a study that used a convolutional neural network (CNN) to detect plant diseases from images with high accuracy. The researchers trained a CNN model on a dataset of plant leaf images labeled with 38 different disease classes. The CNN was able to automatically extract features from the input images and classify them into the respective disease classes. The proposed system achieved an average accuracy of 92%, demonstrating that neural networks can effectively detect plant diseases even with limited computing resources. The document provides details on how CNNs work, including their typical layers of convolution, max pooling, and fully connected layers, and discusses previous related work applying deep learning to plant disease detection.
IRJET- Deep Convolutional Neural Network for Natural Image Matting using Init...IRJET Journal
This document describes a study that used a deep convolutional neural network (CNN) to perform natural image matting using initial alpha mattes. The researchers trained a CNN using alpha mattes generated from closed form matting and K-nearest neighbor (KNN) matting as inputs. Combining the results from these two existing matting methods, which take different local image structures as input, achieved more accurate foreground extraction than prior methods alone. The CNN was able to classify images with higher performance than existing algorithms by using convolutional layers instead of fully connected layers.
This document summarizes a research paper about developing a new set of low-complexity features for detecting steganography in JPEG images. The proposed features, called DCTR features, are computed by taking the discrete cosine transform (DCT) of non-overlapping 8x8 blocks of the image, resulting in 64 feature maps. Histograms are formed from the quantized noise residuals in these feature maps. This approach has lower computational complexity than previous rich models used for steganalysis and provides competitive detection accuracy across different steganographic algorithms while using fewer features. The paper introduces the concept of an undecimated DCT and explains how it relates to previous work in JPEG steganalysis.
Optimized Neural Network for Classification of Multispectral ImagesIDES Editor
This document summarizes an article that proposes using a multiobjective particle swarm optimization (MOPSO) approach to optimize the structure of an artificial neural network for classifying multispectral satellite images. Specifically, the MOPSO is used to simultaneously select the most discriminative spectral bands from the available options and determine the optimal number of nodes in the hidden layer of the neural network. The MOPSO approach is compared to traditional classifiers like maximum likelihood classification and Euclidean classifiers. The results show that the MOPSO-optimized neural network approach provides superior performance for remote sensing image classification problems.
High Speed Data Exchange Algorithm in Telemedicine with Wavelet based on 4D M...Dr. Amarjeet Singh
Existing Medical imaging techniques such as fMRI, positron emission tomography (PET), dynamic 3D ultrasound and dynamic computerized tomography yield large amounts of four-dimensional sets. 4D medical data sets are the series of volumetric images netted in time, large in size and demand a great of assets for storage and transmission. Here, in this paper, we present a method wherein 3D image is taken and Discrete Wavelet Transform(DWT) and Dual-Tree Complex Wavelet Transform(DTCWT) techniques are applied separately on it and the image is split into sub-bands. The encoding and decoding are done using 3D-SPIHT, at different bit per pixels(bpp). The reconstructed image is synthesized using Inverse DWT technique. The quality of the compressed image has been evaluated using some factors such as Mean Square Error(MSE) and Peak-Signal to Noise Ratio (PSNR).
IRJET- Face Recognition using Machine LearningIRJET Journal
This document presents a modified CNN architecture for face recognition that adds two batch normalization operations to improve performance. The CNN extracts facial features using convolutional layers and max pooling, and classifies faces using a softmax classifier. The proposed approach was tested on a face database containing images of 4 individuals with varying lighting conditions. Experimental results showed the modified CNN with batch normalization achieved better recognition results than traditional methods.
A new block cipher for image encryption based on multi chaotic systemsTELKOMNIKA JOURNAL
In this paper, a new algorithm for image encryption is proposed based on three chaotic systems which are Chen system,logistic map and two-dimensional (2D) Arnold cat map. First, a permutation scheme is applied to the image, and then shuffled image is partitioned into blocks of pixels. For each block, Chen system is employed for confusion and then logistic map is employed for generating subsititution-box (S-box) to substitute image blocks. The S-box is dynamic, where it is shuffled for each image block using permutation operation. Then, 2D Arnold cat map is used for providing diffusion, after that XORing the result using Chen system to obtain the encrypted image.The high security of proposed algorithm is experimented using histograms, unified average changing intensity (UACI), number of pixels change rate (NPCR), entropy, correlation and keyspace analyses.
Image Steganography Using Wavelet Transform And Genetic AlgorithmAM Publications
This paper presents the application of Wavelet Transform and Genetic Algorithm in a novel
steganography scheme. We employ a genetic algorithm based mapping function to embed data in Discrete Wavelet
Transform coefficients in 4x4 blocks on the cover image. The optimal pixel adjustment process is applied after
embedding the message. We utilize the frequency domain to improve the robustness of steganography and, we
implement Genetic Algorithm and Optimal Pixel Adjustment Process to obtain an optimal mapping function to
reduce the difference error between the cover and the stego-image, therefore improving the hiding capacity with
low distortions. Our Simulation results reveal that the novel scheme outperforms adaptive steganography technique
based on wavelet transform in terms of peak signal to noise ratio and capacity, 39.94 dB and 50% respectively.
IRJET- Identification of Scene Images using Convolutional Neural Networks - A...IRJET Journal
This document summarizes research on using convolutional neural networks (CNNs) for scene image identification. It first discusses traditional object detection methods and their limitations. CNNs are presented as an improved approach, with convolutional, pooling and fully connected layers to extract features and classify images. Several popular CNN-based object detection algorithms are then summarized, including R-CNN, Fast R-CNN, Faster R-CNN and YOLO. The document concludes that CNN methods provide more accurate object identification compared to traditional algorithms due to their ability to learn from large datasets.
This document is a project report submitted by Shubham Jain and Vikas Jain for their course CS676A. The project aims to learn relative attributes associated with face images using the PubFig dataset. Convolutional neural network features and the RankNet model were used to predict attribute rankings. RankNet achieved better performance than RankSVM and GIST features. Zero-shot learning for unseen classes was explored by building probabilistic class models, but performance was poor. Future work could improve the modeling of unseen classes.
Architecture neural network deep optimizing based on self organizing feature ...journalBEEI
Forward neural network (FNN) execution relying on the algorithm of training and architecture selection. Different parameters using for nip out the architecture of FNN such as the connections number among strata, neurons hidden number in each strata hidden and hidden strata number. Feature architectural combinations exponential could be uncontrollable manually so specific architecture can be design automatically by using special algorithm which build system with ability generalization better. Determination of architecture FNN can be done by using the algorithm of optimization numerous. In this paper methodology new proposes achievement where FNN neurons respective with hidden layers estimation work where in this work collect algorithm training self organizing feature map (SOFM) with advantages to explain how the best architectural selected automatically by SOFM from criteria error testing based on architecture populated. Different size of dataset benchmark of 4 classifications tested for approach proposed.
This document summarizes a study that used neural networks and particle swarm optimization incorporating fuzzy c-means (PSOFCN) segmentation to recognize handwritten characters in the Meetei Mayek script. 34 characters were analyzed. Images were preprocessed, segmented using PSOFCN and recognized using a multilayer feedforward neural network with backpropagation. 1700 samples were used for training and 1700 for testing. Recognition accuracy ranged from 30-100%, with an average of 72%. Characters with simpler shapes had higher accuracy than more complex characters.
International Journal of Research in Engineering and Science is an open access peer-reviewed international forum for scientists involved in research to publish quality and refereed papers. Papers reporting original research or experimentally proved review work are welcome. Papers for publication are selected through peer review to ensure originality, relevance, and readability.
Particle Swarm Optimization for Nano-Particles Extraction from Supporting Mat...CSCJournals
This document summarizes a study that uses Particle Swarm Optimization (PSO) for automatic segmentation of nano-particles in Transmission Electron Microscopy (TEM) images. PSO is applied to specify local and global thresholds for segmentation by treating image entropy as a minimization problem. Results show the PSO method improves over previous techniques by reducing incorrect characterization of nano-particles in images affected by liquid concentrations or supporting materials, with up to a 27% reduction in errors. Compared to manual characterization, PSO provides comparable particle counting with higher computational efficiency suitable for real-time analysis.
The document discusses clustering images based on their properties. Images are converted into intensity, contrast, Weibull and fractal images. Eight properties are calculated for each image type, including brightness, standard deviation, entropy, skewness, kurtosis, separability, spatial frequency and visibility. The properties are normalized and clustered using k-means clustering. Tables show normalized property values for different image types. The clustering groups similar images based on their discriminative properties.
Multilayer extreme learning machine for hand movement prediction based on ele...journalBEEI
Brain computer interface (BCI) technology connects humans with machines via electroencephalography (EEG). The mechanism of BCI is pattern recognition, which proceeds by feature extraction and classification. Various feature extraction and classification methods can differentiate human motor movements, especially those of the hand. Combinations of these methods can greatly improve the accuracy of the results. This article explores the performances of nine feature-extraction types computed by a multilayer extreme learning machine (ML-ELM). The proposed method was tested on different numbers of EEG channels and different ML-ELM structures. Moreover, the performance of ML-ELM was compared with those of ELM, Support Vector Machine and Naive Bayes in classifying real and imaginary hand movements in offline mode. The ML-ELM with discrete wavelet transform (DWT) as feature extraction outperformed the other classification methods with highest accuracy 0.98. So, the authors also found that the structures influenced the accuracy of ML-ELM for different task, feature extraction used and channel used.
This document presents a method for image upscaling using a fuzzy ARTMAP neural network. It begins with an introduction to image upscaling and interpolation techniques. It then provides background on ARTMAP neural networks and fuzzy logic. The proposed method uses a linear interpolation algorithm trained with an ARTMAP network. Results show the method performs better than nearest neighbor interpolation in terms of peak signal-to-noise ratio, mean squared error, and structural similarity, though not as high as bicubic interpolation. Overall, the fuzzy ARTMAP network provides an effective way to perform image upscaling with fewer artifacts than traditional methods.
Semantic Concept Detection in Video Using Hybrid Model of CNN and SVM Classif...CSCJournals
In today's era of digitization and fast internet, many video are uploaded on websites, a mechanism is required to access this video accurately and efficiently. Semantic concept detection achieve this task accurately and is used in many application like multimedia annotation, video summarization, annotation, indexing and retrieval. Video retrieval based on semantic concept is efficient and challenging research area. Semantic concept detection bridges the semantic gap between low level extraction of features from key-frame or shot of video and high level interpretation of the same as semantics. Semantic Concept detection automatically assigns labels to video from predefined vocabulary. This task is considered as supervised machine learning problem. Support vector machine (SVM) emerged as default classifier choice for this task. But recently Deep Convolutional Neural Network (CNN) has shown exceptional performance in this area. CNN requires large dataset for training. In this paper, we present framework for semantic concept detection using hybrid model of SVM and CNN. Global features like color moment, HSV histogram, wavelet transform, grey level co-occurrence matrix and edge orientation histogram are selected as low level features extracted from annotated groundtruth video dataset of TRECVID. In second pipeline, deep features are extracted using pretrained CNN. Dataset is partitioned in three segments to deal with data imbalance issue. Two classifiers are separately trained on all segments and fusion of scores is performed to detect the concepts in test dataset. The system performance is evaluated using Mean Average Precision for multi-label dataset. The performance of the proposed framework using hybrid model of SVM and CNN is comparable to existing approaches.
GeoAI: A Model-Agnostic Meta-Ensemble Zero-Shot Learning Method for Hyperspec...Konstantinos Demertzis
The document discusses a new meta-ensemble zero-shot learning method called MAME-ZsL for hyperspectral image analysis and classification. MAME-ZsL overcomes the difficulties of traditional deep learning methods that require large labeled datasets and long training times. It reduces computational costs, avoids overfitting, and achieves high classification accuracy even when testing classes were not present during training. The method is a novel optimization-based meta-ensemble architecture that facilitates learning representations from limited labeled examples to enable one-shot and zero-shot learning.
We presents a technique for moving objects extraction. There are several different approaches for moving object extraction, clustering is one of object extraction method with a stronger teorical foundation used in many applications. And need high performance in many extraction process of moving object. We compare K-Means and Self-Organizing Map method for extraction moving objects, for performance measurement of moving object extraction by applying MSE and PSNR. According to experimental result that the MSE value of K-Means is smaller than Self-Organizing Map. It is also that PSNR of K-Means is higher than Self-Organizing Map algorithm. The result proves that K-Means is a promising method to cluster pixels in moving objects extraction.
CNN FEATURES ARE ALSO GREAT AT UNSUPERVISED CLASSIFICATION cscpconf
This paper aims at providing insight on the transferability of deep CNN features to
unsupervised problems. We study the impact of different pretrained CNN feature extractors on
the problem of image set clustering for object classification as well as fine-grained
classification. We propose a rather straightforward pipeline combining deep-feature extraction
using a CNN pretrained on ImageNet and a classic clustering algorithm to classify sets of
images. This approach is compared to state-of-the-art algorithms in image-clustering and
provides better results. These results strengthen the belief that supervised training of deep CNN
on large datasets, with a large variability of classes, extracts better features than most carefully
designed engineering approaches, even for unsupervised tasks. We also validate our approach
on a robotic application, consisting in sorting and storing objects smartly based on clustering
Plant Disease Detection using Convolution Neural Network (CNN)IRJET Journal
This document describes a study that used a convolutional neural network (CNN) to detect plant diseases from images with high accuracy. The researchers trained a CNN model on a dataset of plant leaf images labeled with 38 different disease classes. The CNN was able to automatically extract features from the input images and classify them into the respective disease classes. The proposed system achieved an average accuracy of 92%, demonstrating that neural networks can effectively detect plant diseases even with limited computing resources. The document provides details on how CNNs work, including their typical layers of convolution, max pooling, and fully connected layers, and discusses previous related work applying deep learning to plant disease detection.
Hyper-parameter optimization of convolutional neural network based on particl...journalBEEI
The document proposes using a particle swarm optimization (PSO) algorithm to optimize the hyperparameters of a convolutional neural network (CNN) for image classification. The PSO algorithm is used to find optimal values for CNN hyperparameters like the number and size of convolutional filters. In experiments on the MNIST handwritten digit dataset, the optimized CNN achieved a testing error rate of 0.87%, which is competitive with state-of-the-art models. The proposed approach finds optimized CNN architectures automatically without requiring manual design or encoding strategies during training.
The document describes a study that used deep learning and convolutional neural networks to develop an image-based detection model for classifying four types of nuts (hazelnut, walnut, pecan, forest nut) with 100% accuracy on test data. The model was developed using Python in Google Colab, utilizing a dataset of 1595 images. A VGG16 model pre-trained on ImageNet was used to extract features from the images. The model contains convolutional and max pooling layers for feature extraction, and fully connected layers for classification. Training, validation, and testing of the model was performed in Google Colab using a GPU, demonstrating the feasibility of deep learning for nut detection applications.
Classification of Images Using CNN Model and its VariantsIRJET Journal
This document presents research on using convolutional neural network (CNN) models for image classification. It introduces CNNs and describes their basic architecture including convolutional layers, pooling layers, ReLU layers, and fully connected layers. It then discusses implementing three CNN configurations on the CIFAR-10 dataset to classify images into 10 classes. The first model is a simple CNN, while the other two add techniques like dropout regularization to prevent overfitting. The results show dropout regularization can significantly improve accuracy, and lower batch sizes may achieve better results than higher batch sizes. The goal is to compare the CNN variants' performance on image classification.
HANDWRITTEN DIGIT RECOGNITION SYSTEM BASED ON CNN AND SVMsipij
The recognition of handwritten digits has aroused the interest of the scientific community and is the subject
of a large number of research works thanks to its various applications. The objective of this paper is to
develop a system capable of recognizing handwritten digits using a convolutional neural network (CNN)
combined with machine learning approaches to ensure diversity in automatic classification tools. In this
work, we propose a classification method based on deep learning, in particular the convolutional neural
network for feature extraction, it is a powerful tool that has had great success in image classification,
followed by the support vector machine (SVM) for higher performance. We used the dataset (MNIST), and
the results obtained showed that the combination of CNN with SVM improves the performance of the model
as well as the classification accuracy with a rate of 99.12%.
Handwritten Digit Recognition System based on CNN and SVMsipij
This document summarizes a research paper on developing a handwritten digit recognition system using a combination of convolutional neural networks (CNN) and support vector machines (SVM). The proposed system extracts features from images using a CNN, then classifies the images using an SVM. The system was tested on the MNIST dataset and achieved 99.12% classification accuracy, outperforming a basic CNN-only model. This high accuracy demonstrates that combining CNNs with SVMs improves performance for handwritten digit recognition.
Content-based product image retrieval using squared-hinge loss trained convol...IJECEIAES
Convolutional neural networks (CNN) have proven to be highly effective in large-scale object detection and image classification, as well as in serving as feature extractors for content-based image retrieval. While CNN models are typically trained with category label supervision and softmax loss for product image retrieval, we propose a different approach for feature extraction using the squared-hinge loss, an alternative multiclass classification loss function. First, transfer learning is performed on a pre-trained model, followed by finetuning the model. Then, image features are extracted based on the fine-tuned model and indexed using the nearest-neighbor indexing technique. Experiments are conducted on VGG19, InceptionV3, MobileNetV2, and ResNet18 CNN models. The model training results indicate that training the models with squared-hinge loss reduces the loss values in each epoch and reaches stability in less epoch than softmax loss. Retrieval results show that using features from squared-hinge trained models improves the retrieval accuracy by up to 3.7% compared to features from softmax-trained models. Moreover, the squared-hinge trained MobileNetV2 features outperformed others, while the ResNet18 feature gives the advantage of having the lowest dimensionality with competitive accuracy.
Machine learning based augmented reality for improved learning application th...IJECEIAES
Detection of objects and their location in an image are important elements of current research in computer vision. In May 2020, Meta released its state-ofthe-art object-detection model based on a transformer architecture called detection transformer (DETR). There are several object-detection models such as region-based convolutional neural network (R-CNN), you only look once (YOLO) and single shot detectors (SSD), but none have used a transformer to accomplish this task. These models mentioned earlier, use all sorts of hyperparameters and layers. However, the advantages of using a transformer pattern make the architecture simple and easy to implement. In this paper, we determine the name of a chemical experiment through two steps: firstly, by building a DETR model, trained on a customized dataset, and then integrate it into an augmented reality mobile application. By detecting the objects used during the realization of an experiment, we can predict the name of the experiment using a multi-class classification approach. The combination of various computer vision techniques with augmented reality is indeed promising and offers a better user experience.
The document presents a pre-trained image processing transformer (IPT) model that can be applied to multiple low-level computer vision tasks like denoising, super-resolution, and deraining. The IPT model consists of multiple task-specific heads and tails connected to a shared transformer body. The model is trained on over 10 million image pairs generated from the ImageNet dataset by adding synthetic noise and corruption. The input images are encoded into patches and processed by the transformer's encoder-decoder architecture using position and task embeddings. Experimental results show the pre-trained IPT model outperforms state-of-the-art methods on various benchmarks when fine-tuned for specific tasks.
We trained a large, deep convolutional neural network to classify the 1.2 million
high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different
classes. On the test data, we achieved top-1 and top-5 error rates of 37.5%
and 17.0% which is considerably better than the previous state-of-the-art. The
neural network, which has 60 million parameters and 650,000 neurons, consists
of five convolutional layers, some of which are followed by max-pooling layers,
and three fully-connected layers with a final 1000-way softmax. To make training
faster, we used non-saturating neurons and a very efficient GPU implementation
of the convolution operation. To reduce overfitting in the fully-connected
layers we employed a recently-developed regularization method called “dropout”
that proved to be very effective. We also entered a variant of this model in the
ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%,
compared to 26.2% achieved by the second-best entry
The document describes a study that used a convolutional neural network with a ConvNeXtLarge architecture to classify skin cancer images into benign and malignant classes. The CNN model was trained on a dataset of 3,297 skin cancer images from Kaggle. It achieved an AUC of 0.91 for classifying the images, demonstrating the ConvNeXtLarge architecture is effective for this task. The study aims to help early diagnosis and treatment of skin cancers.
IRJET- Art Authentication System using Deep Neural NetworksIRJET Journal
1) The document presents a system to authenticate paintings by artists using deep convolutional neural networks. The system processes images through thousands of neurons to extract patterns and characteristics of an artist's style.
2) A deep convolutional neural network model is implemented and trained on datasets of labeled artworks. The network aims to classify new paintings by artist with 80% accuracy, higher than previous methods.
3) The system was tested on 5 paintings, with a confusion matrix showing correct and incorrect classifications. The 80% accuracy rate is an improvement over previous techniques, but the model has limitations as the number of paintings increases.
Competent scene classification using feature fusion of pre-trained convolutio...TELKOMNIKA JOURNAL
In view of the fact that the development of convolutional neural networks (CNN) and other deep learning techniques, scientists have become more interested in the scene categorization of remotely acquired images as well as other algorithms and datasets. The spatial geometric detail information may be lost as the convolution layer thickness increases, which would have a significant impact on the classification accuracy. Fusion-based techniques, which are regarded to be a viable way to express scene features, have recently attracted a lot of interest as a solution to this issue. Here, we suggested a convolutional feature fusion network that makes use of canonical correlation, which is the linear correlation between two feature maps. Then, to improve scene classification accuracy, the deep features extracted from various pre-trained convolutional neural networks are efficiently fused. We thoroughly evaluated three different fused CNN designs to achieve the best results. Finally, we used the support vector machine for categorization (SVM). In the analysis, two real-world datasets UC Merced and SIRI-WHU were employed, and the competitiveness of the investigated technique was evaluated. The improved categorization accuracy demonstrates that the fusion technique under consideration has produced affirmative results when compared to individual networks.
IRJET-Multiclass Classification Method Based On Deep Learning For Leaf Identi...IRJET Journal
This document discusses a multiclass classification method using deep learning for leaf identification to help farmers. It proposes using a convolutional neural network (CNN) model for feature extraction and classification of leaf images. The CNN model is trained on labeled leaf image data and can then be used to classify new unlabeled leaf images. The method involves preprocessing leaf images, extracting features using the CNN model, and classifying the leaves into different plant categories. The researchers tested their method on 13 plant leaf categories and 4 disease categories, achieving 95.25% accuracy. They conclude CNNs are well-suited for leaf identification and classification tasks due to their ability to handle large image datasets.
Image segmentation and classification tasks in computer vision have proven to be highly effective using neural networks, specifically Convolutional Neural Networks (CNNs). These tasks have numerous
practical applications, such as in medical imaging, autonomous driving, and surveillance. CNNs are capable
of learning complex features directly from images and achieving outstanding performance across several
datasets. In this work, we have utilized three different datasets to investigate the efficacy of various preprocessing and classification techniques in accurssedately segmenting and classifying different structures
within the MRI and natural images. We have utilized both sample gradient and Canny Edge Detection
methods for pre-processing, and K-means clustering have been applied to segment the images. Image
augmentation improves the size and diversity of datasets for training the models for image classification
Image segmentation and classification tasks in computer vision have proven to be highly effective using neural networks, specifically Convolutional Neural Networks (CNNs). These tasks have numerous
practical applications, such as in medical imaging, autonomous driving, and surveillance. CNNs are capable
of learning complex features directly from images and achieving outstanding performance across several
datasets. In this work, we have utilized three different datasets to investigate the efficacy of various preprocessing and classification techniques in accurssedately segmenting and classifying different structures
within the MRI and natural images. We have utilized both sample gradient and Canny Edge Detection
methods for pre-processing, and K-means clustering have been applied to segment the images. Image
augmentation improves the size and diversity of datasets for training the models for image classification.
This work highlights transfer learning’s effectiveness in image classification using CNNs and VGG 16 that
provides insights into the selection of pre-trained models and hyper parameters for optimal performance.
We have proposed a comprehensive approach for image segmentation and classification, incorporating preprocessing techniques, the K-means algorithm for segmentation, and employing deep learning models such
as CNN and VGG 16 for classification.
FaceDetectionforColorImageBasedonMATLAB.pdfAnita Pal
The document proposes a method for face detection in color images using MATLAB and a convolutional neural network (CNN) with two layers. It begins with an introduction to face detection technologies and challenges. It then describes the proposed methodology, which involves reading a color image, applying CNN convolutions to detect faces, and fusing the results. Mathematical aspects of CNNs are discussed, including equations for convolutional operations and dimensions. The CNN method combines feature extraction, segmentation, and classification. Finally, the document summarizes how the proposed algorithm was designed in MATLAB to identify faces using a simple learning algorithm and CNNs with fewer layers than traditional networks. The goal is to develop a fast face detection program for color images based on new deep learning standards.
AUTOMATIC FRUIT RECOGNITION BASED ON DCNN FOR COMMERCIAL SOURCE TRACE SYSTEMijcsa
Automatically fruit recognition by using machine vision is considered as challenging task due to similarities between various types of fruits and external environmental changes e-g lighting. In this paper, fruit recognition algorithm based on Deep Convolution Neural Network(DCNN) is proposed. Most of the previous techniques have some limitations because they were examined and evaluated under limited dataset, furthermore they have not considered external environmental changes. Another major contribution in this paper is that we established fruit images database having 15 different categories comprising of 44406 images which were collected within a period of 6 months by keeping in view the limitations of existing dataset under different real-world conditions. Images were directly used as input to DCNN for training and recognition without extracting features, besides this DCNN learn optimal features from images through adaptation process. The final decision was totally based on a fusion of all regional classification using probability mechanism. Experimental results exhibit that the proposed approach have efficient capability of automatically recognizing the fruit with a high accuracy of 99% and it can also effectively meet real world application requirements.
Similar to TRANSFER LEARNING BASED IMAGE VISUALIZATION USING CNN (20)
artificial intelligence and data science contents.pptxGauravCar
What is artificial intelligence? Artificial intelligence is the ability of a computer or computer-controlled robot to perform tasks that are commonly associated with the intellectual processes characteristic of humans, such as the ability to reason.
› ...
Artificial intelligence (AI) | Definitio
Introduction- e - waste – definition - sources of e-waste– hazardous substances in e-waste - effects of e-waste on environment and human health- need for e-waste management– e-waste handling rules - waste minimization techniques for managing e-waste – recycling of e-waste - disposal treatment methods of e- waste – mechanism of extraction of precious metal from leaching solution-global Scenario of E-waste – E-waste in India- case studies.
Embedded machine learning-based road conditions and driving behavior monitoringIJECEIAES
Car accident rates have increased in recent years, resulting in losses in human lives, properties, and other financial costs. An embedded machine learning-based system is developed to address this critical issue. The system can monitor road conditions, detect driving patterns, and identify aggressive driving behaviors. The system is based on neural networks trained on a comprehensive dataset of driving events, driving styles, and road conditions. The system effectively detects potential risks and helps mitigate the frequency and impact of accidents. The primary goal is to ensure the safety of drivers and vehicles. Collecting data involved gathering information on three key road events: normal street and normal drive, speed bumps, circular yellow speed bumps, and three aggressive driving actions: sudden start, sudden stop, and sudden entry. The gathered data is processed and analyzed using a machine learning system designed for limited power and memory devices. The developed system resulted in 91.9% accuracy, 93.6% precision, and 92% recall. The achieved inference time on an Arduino Nano 33 BLE Sense with a 32-bit CPU running at 64 MHz is 34 ms and requires 2.6 kB peak RAM and 139.9 kB program flash memory, making it suitable for resource-constrained embedded systems.
Software Engineering and Project Management - Introduction, Modeling Concepts...Prakhyath Rai
Introduction, Modeling Concepts and Class Modeling: What is Object orientation? What is OO development? OO Themes; Evidence for usefulness of OO development; OO modeling history. Modeling
as Design technique: Modeling, abstraction, The Three models. Class Modeling: Object and Class Concept, Link and associations concepts, Generalization and Inheritance, A sample class model, Navigation of class models, and UML diagrams
Building the Analysis Models: Requirement Analysis, Analysis Model Approaches, Data modeling Concepts, Object Oriented Analysis, Scenario-Based Modeling, Flow-Oriented Modeling, class Based Modeling, Creating a Behavioral Model.
CHINA’S GEO-ECONOMIC OUTREACH IN CENTRAL ASIAN COUNTRIES AND FUTURE PROSPECTjpsjournal1
The rivalry between prominent international actors for dominance over Central Asia's hydrocarbon
reserves and the ancient silk trade route, along with China's diplomatic endeavours in the area, has been
referred to as the "New Great Game." This research centres on the power struggle, considering
geopolitical, geostrategic, and geoeconomic variables. Topics including trade, political hegemony, oil
politics, and conventional and nontraditional security are all explored and explained by the researcher.
Using Mackinder's Heartland, Spykman Rimland, and Hegemonic Stability theories, examines China's role
in Central Asia. This study adheres to the empirical epistemological method and has taken care of
objectivity. This study analyze primary and secondary research documents critically to elaborate role of
china’s geo economic outreach in central Asian countries and its future prospect. China is thriving in trade,
pipeline politics, and winning states, according to this study, thanks to important instruments like the
Shanghai Cooperation Organisation and the Belt and Road Economic Initiative. According to this study,
China is seeing significant success in commerce, pipeline politics, and gaining influence on other
governments. This success may be attributed to the effective utilisation of key tools such as the Shanghai
Cooperation Organisation and the Belt and Road Economic Initiative.
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Sinan KOZAK
Sinan from the Delivery Hero mobile infrastructure engineering team shares a deep dive into performance acceleration with Gradle build cache optimizations. Sinan shares their journey into solving complex build-cache problems that affect Gradle builds. By understanding the challenges and solutions found in our journey, we aim to demonstrate the possibilities for faster builds. The case study reveals how overlapping outputs and cache misconfigurations led to significant increases in build times, especially as the project scaled up with numerous modules using Paparazzi tests. The journey from diagnosing to defeating cache issues offers invaluable lessons on maintaining cache integrity without sacrificing functionality.
LLM Fine Tuning with QLoRA Cassandra Lunch 4, presented by Anant
TRANSFER LEARNING BASED IMAGE VISUALIZATION USING CNN
1. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
DOI: 10.5121/ijaia.2019.10404 47
TRANSFER LEARNING BASED IMAGE
VISUALIZATION USING CNN
Santosh Giri1
and Basanta Joshi2
1
Department of Computer & Electronics Engineering, Kathford Int'l College of
Engineering and Management, IOE, TU, Nepal
2
Department of Electronics & Computer Engineering, Pulchowk Campus, IOE, TU,
Nepal
ABSTRACT
Image classification is a popular machine learning based applications of deep learning. Deep learning
techniques are very popular because they can be effectively used in performing operations on image data
in large-scale. In this paper CNN model was designed to better classify images. We make use of feature
extraction part of inception v3 model for feature vector calculation and retrained the classification layer
with these feature vector. By using the transfer learning mechanism the classification layer of the CNN
model was trained with 20 classes of Caltech101 image dataset and 17 classes of Oxford 17 flower image
dataset. After training, network was evaluated with testing dataset images from Oxford 17 flower dataset
and Caltech101 image dataset. The mean testing precision of the neural network architecture with
Caltech101 dataset was 98 % and with Oxford 17 Flower image dataset was 92.27 %.
KEYWORDS
Image Classification, CNN, Deep Learning, Transfer Learning.
1. INTRODUCTION
In this modern era, Computers are being powerful day by day. They have become perfect
companion with high speed computing capabilities over the time. Few decades ago it was
believed that machines are only for arithmetic operations but not for complex tasks like speech
recognition, object detection, image classification, language modeling etc. But these days,
situation is inverted. Machines are capable of doing these things more easily with very much
high accuracy. Usual algorithm consisting of finite arithmetic operations cannot provide capacity
to do such complex tasks for machine. For this, Artificial Intelligence provides lots of
techniques. Learning Algorithms are used for such purpose. Huge dataset is required for training
the model with appropriate architecture. Testing is required to evaluate whether the model is
working properly or not. Neural Network is one of AI techniques emerged long ago in 1940s but
technology at that time was not so advanced. It was up at time in 1980s with the development of
back-propagation [1]. Later it was again discarded due to slow learning and expensive
computation. Initially It was believed that only 2 to 3 hidden layers are sufficient for Neural
Network to work properly but later on it is observed that even more layers can represent high
dimensional features of the input signals [2]. Image classification has received extensive
attention since the early years of computer vision research. Classification remains main
problems in image analysis. CNNs are applicable to fields like Speech Recognition[3], text
prediction[4], handwriting generation[5]and so on.
2. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
48
In These days, RAM on a machine is cheap and is available in plenty. we need lots of labeled
data, time and GPU speed to train a CNN model from scratch[6]. With this limitation, it is not
feasible for training the CNN network from scratch, because it is a computationally intensive
task and it takes several days or even weeks with high speed GPU computer[7], which is not
possible with limited resource we have. Defining appropriate model for image classifications
which will produce good result in small training time and minimum CPU speed is the main task
that this paper is intended to do.
2. RELATED WORKS
In [8], Noval general K nearest neighbor classifier GKMNC[9] was used for visual
classification. Sparse representation based method[10] for learning and deriving the weights
coefficients and FISTA[11] was used for optimization. CNN-M[12], a pretrained CNN was used
for image features extractions then marginal PCA[13] is applied to reduce the dimension of the
extracted features. In[14], Alex net[15] model, a deep neural network is used to learn scene
image features. During the training phase, series of transformation such as convolution, max
pooling, etc are performed to obtained image features. Then two classifier SVM[16] classifier
and Softmax[17] classifier are trained using extracted features from the AlexNet model.
In[18], Spatial pyramid pooling was used in CNN to eliminate the fixed size input requirements.
for this new network structure SPP-net was used, which can generate a fixed length
representation regardless of image size. Standard back propagation algorithm[1] was used for
training, regardless of the input image. In[19], Kernalized version of Nave bayes Neabour[20]
was used for image classification and SVM[16] classifier was trained on Bag-Of-Features[21]
for visual classification. In[22], Extension of the HMAX[23], a four level NN has been used for
image classifications. The local filters at first level are integrated into last level complex filters
to provide a exible description of object regions. In[24], Nearest neighbor classifier[20] was
used for visual classifications. SIFT[25] descriptor to describe shape, HSV[26] values to
describe colors and MR filters to describe texture were used.
3. METHODOLOGY
3.1. Image Preprocessing
The learning method used in this experiment is supervised learning[27].In supervised
learning[27], we have to label the data for training & evaluation of the model. For training and
testing the model, Caltech101[28] image dataset and Oxford 17 flower[29] image dataset were
used. In preprocessing, all images from Caltech101[28] and Oxford 17 flower[29] are resized to
299x299X3 because to train a CNN using transfer learning[30], image input size to CNN must
be same as the input size given to original model. Then images from two standard image
dataset[28][29] were divided into training set, validation set and testing set. From caltech101[28]
dataset, amongst 70 images per class, 60 images were used for training & validation and
remaining 10 images were used for testing. For Oxford 17 flower[29] image dataset, amongt 80
images per class, 64 images were used for training & validation and remaining 16 images were
used for testing.
3. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
49
3.2. Cnn Model Design
Fig. 1. Architecture Of Inception V3 Model [6].
The CNN model designed here is based on inception v3[6]. Fig. 1 represent the detail
architecture of Inception model. It is a 42 layer Deep, pretrained CNN model trained on
ImageNet[31], which was 1st runner up in ImageNet Large Scale Visual Recognition
Competition, with lower error rate i.e. top-1 error: 17.2% & top-5 error: 3.58 %. The inception
model has two parts; feature extraction and classification. we make use of the feature extraction
part of inception model and retrained the classification layer with Oxford 17 flower image
dataset[29] and Caltech101 image dataset[28]. To retrain the classification layer we implement
the transfer learning[30]
3.3. Transfer Learning
Inception v3[6] is a convolutional neural network model and by using GPU configured computer
it takes weeks to train from scratch[6], tensorflow[32] a machine learning framework which
provides plateform to train the classification layer with images from Caltech101[28] and Oxford
17 flower[29] using transfer learning mechanism[30]. transfer learning mechanism[30], which
keeps the weights and bias values of the feature extraction layer and removes parameters on
classification layer of inception v3[6]. First the input image of size of 299x299x3 are feded to
feature extraction layer of CNN. After that feature extraction layer calculate the feature values
for each images, feature vector are 2048 float values for each images. then classification layer of
the CNN is trained with these feature vector. The output labels in the classification layer is equal
to the number of image classes on the dataset.
Fig. 2. Transfer Learning Implementation Pipelines.
4. EVALUATION
The pretrained inception model[6] is used here for experimental purpose and the platform used
here is tensorflow[32] and the hardware plateform used here is dell latitude e6410: 2.4 GHZ
processor intel i5, 4 GB RAM. For experiment Oxford 17 flower dataset[29] and Caltech101
dataset[28] were used.
4. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
50
4.1. Dataset
Oxford 17 flower dataset[29]: This dataset consist of 17 category flowers images with 80 images
for each class. The flowers chosen are some common flowers in Britain. The images has been
collected by Maria-Elena Nilsback and Andrew Zisserman of University of Oxford, UK.
Caltech101 image dataset[28]: This dataset has 101 classes of images. It has 40 to 800 images
per class and Collected by Fei-Fei Li, Marco Andreetto, and Marc Aurelio Ranzato in September
2003.
4.2. Evaluation Procedure
To test the model on Caltech dataset[28], 20 classes of images form Caltech101 image
dataset[28] are used. Each category consist total of 70 images. The training and validation set
consist 60 images and testing set consists of 10 images per class. And to test the model on
Oxford 17 flower dataset[29] 17 classes of flower images were used, each class consist of 80
images: 64 images per class for training & validation and 16 images per class for evaluation
were used.
To train the model transfer learning mechanism[30] is implemented on pretrained inception
model[6]. In transfer learning, weights and bias values of the feature extraction layer are kept
same as original CNN model and removes parameters on classification layer of a CNN. To train
the classification layer, training set images from two dataset[29][28] were used. The Back
propagation algorithm[1] was used to train the classification layer of the CNN model and weight
parameters of classification layer is adjusted by using cross entropy cost function by calculating
the cross entropy error between classification layer output and input feature value from feature
extraction layer.
5. RESULTS
5.1. Training
The training accuracy, validation accuracy and cross entropy graph on flower dataset[29] and
Clatech101 dataset[28] are given in Fig. 3, Fig. 4, Fig. 5 and Fig. 6 respectively. The parameters
used for retraining the model are, training steps: 4500, training interval: 1 and learning rate:
0.045. The training and validation accuracy graph in Fig. 3 and Fig. 5 represents accuracy the
CNN model get from training and validation images. And the cross entropy graph in Fig. 4 and
Fig. 6 represents the difference between the actual output and input feature value while training
the Classification layer of CNN model. In Fig. 3, 4, 5, 6 the Blue line represents accuracy &
cross entropy error variation curve on training images and Orange line represents accuracy and
cross entropy error variation curve on validation images from two dataset[29][28]. Fig. 3
represents the training & validation accuracy variation on flower image dataset[29]. Training
accuracy was 91.2% at the beginning of the training process and starts to increase, after
completion of (3/4) training steps it reached and remains to 100%. Validation accuracy was
72.13% during initiation of training and validation process and final validation accuracy was
92.7%. Fig. 4 represents the cross entropy error for training and validation respectively on
flower image[29]. The cross entropy error were 0.58 and 0.69, during initiation of training and
validation process. As the training steps increases the cross entropy error starts to decrease. At
final step of training and validation the training and validation cross entropy error were 0.02 and
0.09.
5. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
51
Fig. 3. Accuracy graph on Oxford 17 flower dataset[29].
Fig. 4. Cross entropy on Oxford 17 flower dataset[29].
Fig. 5. Accuracy graph on Caltech101 dataset[28].
6. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
52
Fig. 6. Cross entropy graph of training on Caltech101 dataset[28].
Fig. 5 represents the training & validation accuracy variation respectively on Caltech101 image
dataset[28]. Training accuracy was 83.6% at the beginning of the training process and starts to
increase, after completion of (3/5) training steps it reached and remains to 100%. Validation
accuracy was 82.7% during initiation of training and validation process and final validation
accuracy was 94.9%. Fig. 6 represents the cross entropy error for training and validation
respectively on Caltech images[28].
The cross entropy were 1.0 and 1.3, during initiation of training and validation process. As the
training steps increases the cross entropy error starts to decrease. At final step of training and
validation process the training and validation cross entropy error were 0.28 and 0.1.
5.2 Testing and Evaluation
The classification result of the system on testing images from flower dataset[29] and Caltech
dataset[28] are represented in evaluation chart given in Fig. 7 and Fig. 8 and the results
comparisons of our system with other experiments are represented in Table 1 and Table 2.
respectively. The 'TP' in the chart is 'True Positive', which represents correctly classified images
by the system.
Fig. 7. Performance evaluation parameters on flower dataset[29].
7. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
53
For testing the model on flower image[29], 16 images were used from each class, i.e. 272
images from 17 classes. Amongst 272 testing images, correctly classified images by the system
were 251 as shown in Fig. 7. For testing the model on caltech image dataset[28], 10 images were
used from
Fig. 8. Performance evaluation metrics on Caltech image dataset [28].
each class, i.e. 200 images from 20 classes. Amongst 200 testing images, correctly classified
images by the system were 196 as shown in Fig. 8.
Table 1. Performance evaluation of our system on flower dataset [29] with other method
S.N. Method Mean Precision
1 [24] 81.3 %
2 [33] 90.4 %
3 Our system 92.27 %
Table 2. Performance evaluation of our system on Caltech101 dataset [28] with other method
S.N. Method Mean Precision
1 [8] 59.12 %
2 [19] 75.2 %
3 [22] 76.32 %
4 [18] 80.4 %
5 Our system 98.0 %
6. CONCLUSIONS
In this paper, the classification layer of pre-trained inception v3 model was re trained
successfully by implementing transfer learning mechanism. The model yields precision of 92.27
% on 17 classes of Oxford 17 flower image dataset and 98.0 % on 20 classes of Caltech101
image dataset.
8. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
54
REFERENCES
[1] Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, and L. D. Jackel,
Backpropagation applied to handwritten zip code recognition," Neural computation, vol. 1, no. 4, pp.
541-551, 1989.
[2 ] A. Krizhevsky, I. Sutskever, and G. E. Hinton, Imagenet classification with deep convolutional
neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
[3] D. R. Reddy, Speech recognition by machine: A review," Proceedings of the IEEE, vol. 64, no. 4,
pp. 501-531, 1976.
[4 A. Coates, B. Carpenter, C. Case, S. Satheesh, B. Suresh, T. Wang, D. J. Wu, and A. Y. Ng, Text
detection and character recognition in scene images with unsupervised feature learning,"in
Document Analysis and Recognition (ICDAR), 2011 International Conference on. IEEE, 2011, pp.
440-445.
[5] T. Varga, D. Kilchhofer, and H. Bunke, Template-based synthetic handwriting generation for the
training of recognition systems," in Proceedings of the 12th Conference of the International
Graphonomics Society, 2005, pp. 206-211.
[6] C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, Rethinking the inception architecture
for computer vision," in Proceedings of the IEEE conference on computer vision and pattern
recognition, 2016, pp. 2818-2826.
[7] C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A.
Rabinovich, Going deeper with convolutions," in Proceedings of the IEEE conference on computer
vision and pattern recognition, 2015, pp. 1-9.
[8] Q. Liu, A. Puthenputhussery, and C. Liu, Novel general knn classifier and general nearest mean
classifier for visual classification," in Image Processing (ICIP), 2015 IEEE International Conference
on. IEEE, 2015, pp. 1810-1814.
[9] J. M. Keller, M. R. Gray, and J. A. Givens, A fuzzy k-nearest neighbor algorithm," IEEE
transactions on systems, man, and cybernetics, no. 4, pp. 580-585, 1985.
[10] J. A. Tropp, Greed is good: Algorithmic results for sparse approximation," IEEE Transactions on
Information theory, vol. 50, no. 10, pp. 2231-2242, 2004.
[11] A. Beck and M. Teboulle, A fast iterative shrinkage-thresholding algorithm for linear inverse
problems," SIAM journal on imaging sciences, vol. 2, no. 1, pp. 183-202, 2009.
[12] K. Chatfield, K. Simonyan, A. Vedaldi, and A. Zisserman, Return of the devil in the details: Delving
deep into convolutional nets," arXiv preprint arXiv:1405.3531, 2014.
[13] I. Jolliffe, Principal component analysis," in International encyclopedia of statistical science.
Springer, 2011, pp. 1094-1096.
[14] J. Sun, X. Cai, F. Sun, and J. Zhang, Scene image classification method based on alex-net model," in
Informative and Cybernetics for Computational Social Systems (ICCSS), 2016 3rd International
Conference on. IEEE, 2016, pp. 363-367.
[15] A. Krizhevsky, I. Sutskever, and G. E. Hinton, Imagenet classification with deep convolutional
neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
[16] C. Cortes and V. Vapnik, Support-vector networks," Machine learning, vol. 20, no. 3, pp. 273-297,
1995.
9. International Journal of Artificial Intelligence and Applications (IJAIA), Vol.10, No.4, July 2019
55
[17] N. M. Nasrabadi, Pattern recognition and machine learning," Journal of electronic imaging, vol. 16,
no. 4, p. 049901, 2007.
[18] K. He, X. Zhang, S. Ren, and J. Sun, Spatial pyramid pooling in deep convolutional networks for
visual recognition," in European conference on computer vision.Springer, 2014, pp. 346-361.
[19] T. Tuytelaars, M. Fritz, K. Saenko, and T. Darrell, The nbnn kernel," in Computer Vision (ICCV),
2011 IEEE International Conference on. IEEE, 2011, pp. 1824-1831.
[20] K. P. Murphy et al., Naive bayes classifiers," University of British Columbia, vol. 18, 2006.
[21] Z. S. Harris, Distributional structure," Word, vol. 10, no. 2-3, pp. 146-162, 1954.
[22] C. Theriault, N. Thome, and M. Cord, Extended coding and pooling in the hmax model," IEEE
Transactions on Image Processing, vol. 22, no. 2, pp. 764-777, 2013.
[23] M. Riesenhuber and T. Poggio, Hierarchical models of object recognition in cortex," Nature
neuroscience, vol. 2, no. 11, p. 1019, 1999.
[24] M.-E. Nilsback and A. Zisserman, A visual vocabulary for flower classification," in Computer
Vision and Pattern Recognition, 2006 IEEE Computer Society Conference on, vol. 2. IEEE, 2006,
pp. 1447-1454.
[25] D. G. Lowe, Object recognition from local scale-invariant features," in Computer vision, 1999. The
proceedings of the seventh IEEE international conference on, vol. 2. Ieee, 1999, pp. 1150-1157.
[26] A. R. Smith, Color gamut transform pairs," ACM Siggraph Computer Graphics, vol. 12, no. 3, pp.
12-19, 1978.
[27] S. B. Kotsiantis, I. Zaharakis, and P. Pintelas, Supervised machine learning: A review of
classification techniques," Emerging artificial intelligence applications in computer engineering, vol.
160, pp. 3-24, 2007.
[28] L. Fei-Fei, R. Fergus, and P. Perona, Learning generative visual models from few training examples:
An incremental bayesian approach tested on 101 object categories," Computer vision and Image
understanding, vol. 106, no. 1, pp. 59-70, 2007.
[29] M.-E. Nilsback and A. Zisserman, A visual vocabulary for flower classification," in Computer
Vision and Pattern Recognition, 2006 IEEE Computer Society Conference on, vol. 2. IEEE, 2006,
pp. 1447-1454.
[30] S. J. Pan, Q. Yang et al., A survey on transfer learning," IEEE Transactions on knowledge and data
engineering, vol. 22, no. 10, pp. 1345-1359, 2010.
[31] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, ImageNet: A Large-Scale Hierarchical
Image Database," in CVPR09, 2009.
[32] M. Abadi, P. Barham, J. Chen, Z. Chen, A. Davis, J. Dean, M. Devin, S. Ghemawat, G. Irving, M.
Isard et al., Tensor flow: a system for large-scale machine learning." in OSDI, vol. 16, 2016, pp.
265-283.
[33] Y. Chai, V. Lempitsky, and A. Zisserman, Bicos: A bi-level co-segmentation method for image
classification," 2011.