This document proposes a neuro-fuzzy based classification method called NFMResnet for classifying brain tumor images. NFMResnet combines a MResNet convolutional neural network with a fuzzy self-organization layer. It involves three steps for training: feature extraction using GLCM and shape descriptors, feature selection using the Genboruta algorithm, and fuzzy-based classification. The fuzzy-based classification uses fuzzy c-means clustering to assign membership degrees between clusters rather than single classifications. This allows for more accurate representation of complex or overlapping data. The document finds that combining MResNet with fuzzy logic through NFMResnet increases classification accuracy for brain tumor images.
Hybrid model for detection of brain tumor using convolution neural networksCSITiaesprime
The development of aberrant brain cells, some of which may turn cancerous, is known as a brain tumor. Magnetic resonance imaging (MRI) scans are the most common technique for finding brain tumors. Information about the aberrant tissue growth in the brain is discernible from the MRI scans. In numerous research papers, machine learning, and deep learning algorithms are used to detect brain tumors. It takes extremely little time to forecast a brain tumor when these algorithms are applied to MRI pictures, and better accuracy makes it easier to treat patients. The radiologist can make speedy decisions because of this forecast. The proposed work creates a hybrid convolution neural networks (CNN) model using CNN for feature extraction and logistic regression (LR). The pre-trained model visual geometry group 16 (VGG16) is used for the extraction of features. To reduce the complexity and parameters to train we eliminated the last eight layers of VGG16. From this transformed model the features are extracted in the form of a vector array. These features fed into different machine learning classifiers like support vector machine (SVM), naïve bayes (NB), LR, extreme gradient boosting (XGBoost), AdaBoost, and random forest for training and testing. The performance of different classifiers is compared. The CNN-LR hybrid combination outperformed the remaining classifiers. The evaluation measures such as recall, precision, F1-score, and accuracy of the proposed CNN-LR model are 94%, 94%, 94%, and 91% respectively.
Survey of various methods used for integrating machine learning into brain tu...Drjabez
This document surveys various machine learning methods used for integrating machine learning into brain tumor detection and classification from MRI images. It discusses preprocessing techniques like median filtering, Gaussian high pass filtering, and morphology dilation to enhance images. Segmentation techniques covered include thresholding, edge detection, region-based, watershed, Berkeley wavelet transform, K-means clustering, and neural networks. Feature extraction calculates correlation, skewness. Classification algorithms discussed are multi-layer perceptron, naive Bayes, and support vector machines. The document provides an overview of key steps and methods for machine learning-based brain tumor detection and segmentation from MRI images.
This document describes a project report submitted by three students for their Bachelor of Engineering degree. The project involves developing a system for classifying brain images using machine learning techniques. It discusses challenges in detecting brain tumors and the need for automated classification methods. It also provides an overview of techniques for image segmentation, clustering, and feature extraction that will be used in the project.
SWARM OPTIMIZED MODULAR NEURAL NETWORK BASED DIAGNOSTIC SYSTEM FOR BREAST CAN...ijscai
The document describes a modular neural network approach optimized by particle swarm optimization for breast cancer diagnosis. The approach uses a modular neural network with several independent neural network experts that analyze input data individually and provide outputs that are combined by an integrator. Particle swarm optimization is used to determine optimal connections for each expert neural network during training. The optimized modular neural network is then used to classify breast cancer samples as cancerous or non-cancerous, demonstrating better diagnostic ability than traditional methods.
This document discusses using convolutional neural networks (CNNs) for brain tumor detection. CNNs have convolutional, pooling, and dense layers that can extract features from medical images like MRIs. The document focuses on a CNN model to classify brain images as either containing a tumor or being healthy. The model is trained on a dataset of 259 brain tumor images and 20 healthy images from Kaggle, with data augmentation used to address class imbalance. The CNN contains convolution, pooling, dropout and dense layers with rectified linear unit activation. It aims to accurately detect tumors in brain MRI images.
USING SINGULAR VALUE DECOMPOSITION IN A CONVOLUTIONAL NEURAL NETWORK TO IMPRO...ijcsit
A brain tumor consists of cells showing abnormal brain growth. The area of the brain tumor significantly
affects choosing the type of treatment and following the course of the disease during the treatment. At the
same time, pictures of Brain MRIs are accompanied by noise. Eliminating existing noises can significantly
impact the better segmentation and diagnosis of brain tumors. In this work, we have tried using the
analysis of eigenvalues. We have used the MSVD algorithm, reducing the image noise and then using the
deep neural network to segment the tumor in the images. The proposed method's accuracy was increased
by 2.4% compared to using the original images. With Using the MSVD method, convergence speed has
also increased, showing the proposed method's effectiveness.
The document discusses using singular value decomposition (SVD) to reduce noise in MRI images before using a convolutional neural network (CNN) for brain tumor segmentation. SVD is applied using multiresolution SVD (MSVD) to decompose images into sub-bands and remove noise from high-frequency sub-bands. A U-Net CNN is then used to segment tumors. Results found MSVD improved segmentation accuracy by 2.4% over original images and increased CNN convergence speed. The proposed method effectively combined MSVD denoising with CNN segmentation for improved and faster brain tumor detection.
11.artificial neural network based cancer cell classificationAlexander Decker
This summary provides the high level information from the document in 3 sentences:
The document presents an artificial neural network (ANN) based system called ANN-C3 for cancer cell classification using medical pathological images. ANN-C3 performs image pre-processing, segmentation using Harris corner detection and region growing, feature extraction of Tamura texture features, and classification of cells using a neural network. The system was able to accurately segment and classify cancerous versus non-cancerous cells in pathological images when compared to manual methods.
Hybrid model for detection of brain tumor using convolution neural networksCSITiaesprime
The development of aberrant brain cells, some of which may turn cancerous, is known as a brain tumor. Magnetic resonance imaging (MRI) scans are the most common technique for finding brain tumors. Information about the aberrant tissue growth in the brain is discernible from the MRI scans. In numerous research papers, machine learning, and deep learning algorithms are used to detect brain tumors. It takes extremely little time to forecast a brain tumor when these algorithms are applied to MRI pictures, and better accuracy makes it easier to treat patients. The radiologist can make speedy decisions because of this forecast. The proposed work creates a hybrid convolution neural networks (CNN) model using CNN for feature extraction and logistic regression (LR). The pre-trained model visual geometry group 16 (VGG16) is used for the extraction of features. To reduce the complexity and parameters to train we eliminated the last eight layers of VGG16. From this transformed model the features are extracted in the form of a vector array. These features fed into different machine learning classifiers like support vector machine (SVM), naïve bayes (NB), LR, extreme gradient boosting (XGBoost), AdaBoost, and random forest for training and testing. The performance of different classifiers is compared. The CNN-LR hybrid combination outperformed the remaining classifiers. The evaluation measures such as recall, precision, F1-score, and accuracy of the proposed CNN-LR model are 94%, 94%, 94%, and 91% respectively.
Survey of various methods used for integrating machine learning into brain tu...Drjabez
This document surveys various machine learning methods used for integrating machine learning into brain tumor detection and classification from MRI images. It discusses preprocessing techniques like median filtering, Gaussian high pass filtering, and morphology dilation to enhance images. Segmentation techniques covered include thresholding, edge detection, region-based, watershed, Berkeley wavelet transform, K-means clustering, and neural networks. Feature extraction calculates correlation, skewness. Classification algorithms discussed are multi-layer perceptron, naive Bayes, and support vector machines. The document provides an overview of key steps and methods for machine learning-based brain tumor detection and segmentation from MRI images.
This document describes a project report submitted by three students for their Bachelor of Engineering degree. The project involves developing a system for classifying brain images using machine learning techniques. It discusses challenges in detecting brain tumors and the need for automated classification methods. It also provides an overview of techniques for image segmentation, clustering, and feature extraction that will be used in the project.
SWARM OPTIMIZED MODULAR NEURAL NETWORK BASED DIAGNOSTIC SYSTEM FOR BREAST CAN...ijscai
The document describes a modular neural network approach optimized by particle swarm optimization for breast cancer diagnosis. The approach uses a modular neural network with several independent neural network experts that analyze input data individually and provide outputs that are combined by an integrator. Particle swarm optimization is used to determine optimal connections for each expert neural network during training. The optimized modular neural network is then used to classify breast cancer samples as cancerous or non-cancerous, demonstrating better diagnostic ability than traditional methods.
This document discusses using convolutional neural networks (CNNs) for brain tumor detection. CNNs have convolutional, pooling, and dense layers that can extract features from medical images like MRIs. The document focuses on a CNN model to classify brain images as either containing a tumor or being healthy. The model is trained on a dataset of 259 brain tumor images and 20 healthy images from Kaggle, with data augmentation used to address class imbalance. The CNN contains convolution, pooling, dropout and dense layers with rectified linear unit activation. It aims to accurately detect tumors in brain MRI images.
USING SINGULAR VALUE DECOMPOSITION IN A CONVOLUTIONAL NEURAL NETWORK TO IMPRO...ijcsit
A brain tumor consists of cells showing abnormal brain growth. The area of the brain tumor significantly
affects choosing the type of treatment and following the course of the disease during the treatment. At the
same time, pictures of Brain MRIs are accompanied by noise. Eliminating existing noises can significantly
impact the better segmentation and diagnosis of brain tumors. In this work, we have tried using the
analysis of eigenvalues. We have used the MSVD algorithm, reducing the image noise and then using the
deep neural network to segment the tumor in the images. The proposed method's accuracy was increased
by 2.4% compared to using the original images. With Using the MSVD method, convergence speed has
also increased, showing the proposed method's effectiveness.
The document discusses using singular value decomposition (SVD) to reduce noise in MRI images before using a convolutional neural network (CNN) for brain tumor segmentation. SVD is applied using multiresolution SVD (MSVD) to decompose images into sub-bands and remove noise from high-frequency sub-bands. A U-Net CNN is then used to segment tumors. Results found MSVD improved segmentation accuracy by 2.4% over original images and increased CNN convergence speed. The proposed method effectively combined MSVD denoising with CNN segmentation for improved and faster brain tumor detection.
11.artificial neural network based cancer cell classificationAlexander Decker
This summary provides the high level information from the document in 3 sentences:
The document presents an artificial neural network (ANN) based system called ANN-C3 for cancer cell classification using medical pathological images. ANN-C3 performs image pre-processing, segmentation using Harris corner detection and region growing, feature extraction of Tamura texture features, and classification of cells using a neural network. The system was able to accurately segment and classify cancerous versus non-cancerous cells in pathological images when compared to manual methods.
Artificial neural network based cancer cell classificationAlexander Decker
This document summarizes an artificial neural network (ANN) based system called ANN-C3 for cancer cell classification using medical images. The system performs image pre-processing, segmentation using Harris corner detection and region growing, feature extraction of Tamura texture features, and classification using a neural network ensemble. Segmentation detects threshold points using Harris corner detection and performs region growing from these seed points. Feature extraction converts the image data into numerical form using Tamura texture features that capture variations in illumination and surfaces that human vision and surgeons use to differentiate cancerous and non-cancerous cells. The neural network is trained on a large set of labeled data to accurately classify cells.
Feature Extraction for Image Classification and Analysis with Ant Colony Opti...sipij
The problem of structure extraction from the image which contains many clustered objects is a challenging one for high level image analysis. When an image contains many clustered objects overlapping of objects can cause for hiding the structure. The existing segmentation techniques for better understanding, not able to the address the constituent parts of the image implicitly. The approaches like multistage segmentation address to some extent, but for each stage a separate structure is extracted, and thus causes for the ambiguity about the structure. The proposed approach called Ant Colony Optimization and Fuzzy logic based technique resolves this problem, and gives the implicit structure, that meets with original structure. The segmentation approach uses the swarm intelligence technique based on the behavior of the ant colonies. The segmentation is the process of separating the non-overlapping regions that constitute an image. The segmentation is important for structured and non-structured image analysis and classification for better understanding.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
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.
This document discusses applying a neural network approach to decision making in a self-organizing computing network (SOCN). It proposes using concepts from fuzzy logic and neural networks to build a computing network that can handle mixed data types, like symbolic and numeric data. The network would have input, hidden, and output layers connected by transfer functions. The hidden cells would self-organize based on training data to learn relationships between input and output cells. This approach aims to allow the network to make decisions on data sets with diverse attribute types in a more effective way than other techniques.
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.
3D Segmentation of Brain Tumor ImagingIJAEMSJORNAL
A brain tumor is a collection of anomalous cells that grow in or around the brain. Brain tumors affect the humans badly, it can disrupt proper brain function and be life-threatening. In this project, we have proposed a system to detect, segment, and classify the tumors present in the brain. Once the brain tumor is identified at the very beginning, proper treatments can be done and it may be cured.
Discover How Scientific Data is Used for the Public Good with Natural Languag...BaoTramDuong2
This document discusses using natural language processing techniques like n-grams, deep learning models, and named entity recognition to analyze scientific publications and identify references to datasets. It evaluates classifiers like recurrent neural networks and convolutional neural networks to perform sequence labeling and extract dataset citations. The goal is to help government agencies and researchers quickly find datasets, measures, and experts by automating the analysis of research articles.
Automated Analysis of Microscopy Images using Deep Convolutional Neural NetworkAdetayoOkunoye
This document summarizes research on using deep convolutional neural networks to automatically analyze microscopy images. The goals are to expedite the analysis of high-content microscopy data and automate tasks like cell counting and classification. The researchers trained and tested models using TensorFlow on microscopy images to classify cells, achieving over 75% accuracy. This level of automation could benefit biological research by reducing human errors and speeding up analysis of large image datasets.
A Parallel Framework For Multilayer Perceptron For Human Face RecognitionCSCJournals
Artificial neural networks have already shown their success in face recognition and similar complex pattern recognition tasks. However, a major disadvantage of the technique is that it is extremely slow during training for larger classes and hence not suitable for real-time complex problems such as pattern recognition. This is an attempt to develop a parallel framework for the training algorithm of a perceptron. In this paper, two general architectures for a Multilayer Perceptron (MLP) have been demonstrated. The first architecture is All-Class-in-One-Network (ACON) where all the classes are placed in a single network and the second one is One-Class-in-One-Network (OCON) where an individual single network is responsible for each and every class. Capabilities of these two architectures were compared and verified in solving human face recognition, which is a complex pattern recognition task where several factors affect the recognition performance like pose variations, facial expression changes, occlusions, and most importantly illumination changes. Experimental results show that the proposed OCON structure performs better than the conventional ACON in terms of network training convergence speed and which can be easily exercised in a parallel environment.
Ijeee 16-19-a novel approach to brain tumor classification using wavelet and ...Kumar Goud
This document presents a novel approach for classifying brain tumors using magnetic resonance images (MRIs). The proposed technique uses two stages: 1) discrete wavelet transform for dimensionality reduction and feature extraction, and 2) probabilistic neural network (PNN) for classification. MRIs of benign and malignant brain tumors were collected and preprocessed using discrete wavelet transform to extract features. A PNN classifier was then trained on these features to classify tumors as benign or malignant. The technique aims to provide an automated brain tumor classification method using artificial intelligence.
A review on detecting brain tumors using deep learning and magnetic resonanc...IJECEIAES
Early detection and treatment in the medical field offer a critical opportunity to survive people. However, the brain has a significant role in human life as it handles most human body activities. Accurate diagnosis of brain tumors dramatically helps speed up the patient's recovery and the cost of treatment. Magnetic resonance imaging (MRI) is a commonly used technique due to the massive progress of artificial intelligence in medicine, machine learning, and recently, deep learning has shown significant results in detecting brain tumors. This review paper is a comprehensive article suitable as a starting point for researchers to demonstrate essential aspects of using deep learning in diagnosing brain tumors. More specifically, it has been restricted to only detecting brain tumors (binary classification as normal or tumor) using MRI datasets in 2020 and 2021. In addition, the paper presents the frequently used datasets, convolutional neural network architectures (standard and designed), and transfer learning techniques. The crucial limitations of applying the deep learning approach, including a lack of datasets, overfitting, and vanishing gradient problems, are also discussed. Finally, alternative solutions for these limitations are obtained.
Evaluation of deep neural network architectures in the identification of bone...TELKOMNIKA JOURNAL
This document evaluates the performance of three deep neural network architectures - ResNet, DenseNet, and NASNet - in identifying bone fissures in radiological images. The networks were trained on a dataset of 1000 labeled images of fissured and seamless bones. NASNet achieved the best performance with 75% accuracy, outperforming ResNet and DenseNet. While all networks reduced classification errors, NASNet did so with the fewest parameters. The document concludes NASNet is the best solution for this bone fissure identification task.
Toward enhancement of deep learning techniques using fuzzy logic: a survey IJECEIAES
This document provides an overview of deep learning techniques and how fuzzy logic can be used to enhance them. It discusses how deep learning works and some of its applications, such as self-driving cars, sentiment analysis, virtual assistants, and healthcare. It also provides an introduction to fuzzy logic and how it can simulate human thinking better than binary logic by allowing for degrees of truth. The document surveys previous studies that have combined deep learning and fuzzy logic models to improve deep learning performance by making the models better able to handle imprecise or ambiguous real-world data.
PADDY CROP DISEASE DETECTION USING SVM AND CNN ALGORITHMIRJET Journal
- The document discusses a study on detecting diseases in paddy/rice crops using deep learning algorithms like convolutional neural networks (CNN) and support vector machines (SVM).
- A dataset of rice leaf images was created and a CNN model using transfer learning with MobileNet was developed and trained on the dataset to classify rice diseases.
- The proposed method aims to automatically classify rice disease images to help farmers more accurately identify diseases, as manual identification can be difficult and inaccurate. This could help improve treatment and support farmers.
Artificial neural networks (ANNs) are computing systems inspired by biological neural networks. ANNs can learn complex patterns and make predictions based on large amounts of data. The document discusses the basic structure and functioning of ANNs, including their ability to learn through adjustment of synaptic weights between neurons. It also describes several common types of ANNs, focusing on perceptrons and multi-layer perceptrons.
Artificial neural networks (ANNs) are computing systems inspired by biological neural networks. ANNs can learn complex patterns and make predictions based on large amounts of data. The document discusses the basic structure and functioning of ANNs, including their ability to learn through adjustment of synaptic weights between neurons. It also describes several common types of ANNs, focusing on perceptrons and multi-layer perceptrons.
Pattern Recognition using Artificial Neural NetworkEditor IJCATR
An artificial neural network (ANN) usually called neural network. It can be considered as a resemblance to a paradigm
which is inspired by biological nervous system. In network the signals are transmitted by the means of connections links. The links
possess an associated way which is multiplied along with the incoming signal. The output signal is obtained by applying activation to
the net input NN are one of the most exciting and challenging research areas. As ANN mature into commercial systems, they are likely
to be implemented in hardware. Their fault tolerance and reliability are therefore vital to the functioning of the system in which they
are embedded. The pattern recognition system is implemented with Back propagation network and Hopfield network to remove the
distortion from the input. The Hopfield network has high fault tolerance which supports this system to get the accurate output.
Deep learning is a collection of machine learning algorithms utilizing multiple layers, with which higher levels of raw data are slowly removed. For example, lower layers can recognize edges in image processing whereas higher layers may define concepts for humans such as numbers or letters or faces. In this paper we have done a literature survey of some other papers to know how useful is Deep Learning and how to define other Artificial Intelligence things using Deep Learning. Anirban Chakraborty "A Study of Deep Learning Applications" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-4 , June 2020, URL: https://www.ijtsrd.com/papers/ijtsrd31629.pdf Paper Url :https://www.ijtsrd.com/computer-science/artificial-intelligence/31629/a-study-of-deep-learning-applications/anirban-chakraborty
Artificial neural network based cancer cell classificationAlexander Decker
This document summarizes an artificial neural network (ANN) based system called ANN-C3 for cancer cell classification using medical images. The system performs image pre-processing, segmentation using Harris corner detection and region growing, feature extraction of Tamura texture features, and classification using a neural network ensemble. Segmentation detects threshold points using Harris corner detection and performs region growing from these seed points. Feature extraction converts the image data into numerical form using Tamura texture features that capture variations in illumination and surfaces that human vision and surgeons use to differentiate cancerous and non-cancerous cells. The neural network is trained on a large set of labeled data to accurately classify cells.
Feature Extraction for Image Classification and Analysis with Ant Colony Opti...sipij
The problem of structure extraction from the image which contains many clustered objects is a challenging one for high level image analysis. When an image contains many clustered objects overlapping of objects can cause for hiding the structure. The existing segmentation techniques for better understanding, not able to the address the constituent parts of the image implicitly. The approaches like multistage segmentation address to some extent, but for each stage a separate structure is extracted, and thus causes for the ambiguity about the structure. The proposed approach called Ant Colony Optimization and Fuzzy logic based technique resolves this problem, and gives the implicit structure, that meets with original structure. The segmentation approach uses the swarm intelligence technique based on the behavior of the ant colonies. The segmentation is the process of separating the non-overlapping regions that constitute an image. The segmentation is important for structured and non-structured image analysis and classification for better understanding.
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
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.
This document discusses applying a neural network approach to decision making in a self-organizing computing network (SOCN). It proposes using concepts from fuzzy logic and neural networks to build a computing network that can handle mixed data types, like symbolic and numeric data. The network would have input, hidden, and output layers connected by transfer functions. The hidden cells would self-organize based on training data to learn relationships between input and output cells. This approach aims to allow the network to make decisions on data sets with diverse attribute types in a more effective way than other techniques.
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.
3D Segmentation of Brain Tumor ImagingIJAEMSJORNAL
A brain tumor is a collection of anomalous cells that grow in or around the brain. Brain tumors affect the humans badly, it can disrupt proper brain function and be life-threatening. In this project, we have proposed a system to detect, segment, and classify the tumors present in the brain. Once the brain tumor is identified at the very beginning, proper treatments can be done and it may be cured.
Discover How Scientific Data is Used for the Public Good with Natural Languag...BaoTramDuong2
This document discusses using natural language processing techniques like n-grams, deep learning models, and named entity recognition to analyze scientific publications and identify references to datasets. It evaluates classifiers like recurrent neural networks and convolutional neural networks to perform sequence labeling and extract dataset citations. The goal is to help government agencies and researchers quickly find datasets, measures, and experts by automating the analysis of research articles.
Automated Analysis of Microscopy Images using Deep Convolutional Neural NetworkAdetayoOkunoye
This document summarizes research on using deep convolutional neural networks to automatically analyze microscopy images. The goals are to expedite the analysis of high-content microscopy data and automate tasks like cell counting and classification. The researchers trained and tested models using TensorFlow on microscopy images to classify cells, achieving over 75% accuracy. This level of automation could benefit biological research by reducing human errors and speeding up analysis of large image datasets.
A Parallel Framework For Multilayer Perceptron For Human Face RecognitionCSCJournals
Artificial neural networks have already shown their success in face recognition and similar complex pattern recognition tasks. However, a major disadvantage of the technique is that it is extremely slow during training for larger classes and hence not suitable for real-time complex problems such as pattern recognition. This is an attempt to develop a parallel framework for the training algorithm of a perceptron. In this paper, two general architectures for a Multilayer Perceptron (MLP) have been demonstrated. The first architecture is All-Class-in-One-Network (ACON) where all the classes are placed in a single network and the second one is One-Class-in-One-Network (OCON) where an individual single network is responsible for each and every class. Capabilities of these two architectures were compared and verified in solving human face recognition, which is a complex pattern recognition task where several factors affect the recognition performance like pose variations, facial expression changes, occlusions, and most importantly illumination changes. Experimental results show that the proposed OCON structure performs better than the conventional ACON in terms of network training convergence speed and which can be easily exercised in a parallel environment.
Ijeee 16-19-a novel approach to brain tumor classification using wavelet and ...Kumar Goud
This document presents a novel approach for classifying brain tumors using magnetic resonance images (MRIs). The proposed technique uses two stages: 1) discrete wavelet transform for dimensionality reduction and feature extraction, and 2) probabilistic neural network (PNN) for classification. MRIs of benign and malignant brain tumors were collected and preprocessed using discrete wavelet transform to extract features. A PNN classifier was then trained on these features to classify tumors as benign or malignant. The technique aims to provide an automated brain tumor classification method using artificial intelligence.
A review on detecting brain tumors using deep learning and magnetic resonanc...IJECEIAES
Early detection and treatment in the medical field offer a critical opportunity to survive people. However, the brain has a significant role in human life as it handles most human body activities. Accurate diagnosis of brain tumors dramatically helps speed up the patient's recovery and the cost of treatment. Magnetic resonance imaging (MRI) is a commonly used technique due to the massive progress of artificial intelligence in medicine, machine learning, and recently, deep learning has shown significant results in detecting brain tumors. This review paper is a comprehensive article suitable as a starting point for researchers to demonstrate essential aspects of using deep learning in diagnosing brain tumors. More specifically, it has been restricted to only detecting brain tumors (binary classification as normal or tumor) using MRI datasets in 2020 and 2021. In addition, the paper presents the frequently used datasets, convolutional neural network architectures (standard and designed), and transfer learning techniques. The crucial limitations of applying the deep learning approach, including a lack of datasets, overfitting, and vanishing gradient problems, are also discussed. Finally, alternative solutions for these limitations are obtained.
Evaluation of deep neural network architectures in the identification of bone...TELKOMNIKA JOURNAL
This document evaluates the performance of three deep neural network architectures - ResNet, DenseNet, and NASNet - in identifying bone fissures in radiological images. The networks were trained on a dataset of 1000 labeled images of fissured and seamless bones. NASNet achieved the best performance with 75% accuracy, outperforming ResNet and DenseNet. While all networks reduced classification errors, NASNet did so with the fewest parameters. The document concludes NASNet is the best solution for this bone fissure identification task.
Toward enhancement of deep learning techniques using fuzzy logic: a survey IJECEIAES
This document provides an overview of deep learning techniques and how fuzzy logic can be used to enhance them. It discusses how deep learning works and some of its applications, such as self-driving cars, sentiment analysis, virtual assistants, and healthcare. It also provides an introduction to fuzzy logic and how it can simulate human thinking better than binary logic by allowing for degrees of truth. The document surveys previous studies that have combined deep learning and fuzzy logic models to improve deep learning performance by making the models better able to handle imprecise or ambiguous real-world data.
PADDY CROP DISEASE DETECTION USING SVM AND CNN ALGORITHMIRJET Journal
- The document discusses a study on detecting diseases in paddy/rice crops using deep learning algorithms like convolutional neural networks (CNN) and support vector machines (SVM).
- A dataset of rice leaf images was created and a CNN model using transfer learning with MobileNet was developed and trained on the dataset to classify rice diseases.
- The proposed method aims to automatically classify rice disease images to help farmers more accurately identify diseases, as manual identification can be difficult and inaccurate. This could help improve treatment and support farmers.
Artificial neural networks (ANNs) are computing systems inspired by biological neural networks. ANNs can learn complex patterns and make predictions based on large amounts of data. The document discusses the basic structure and functioning of ANNs, including their ability to learn through adjustment of synaptic weights between neurons. It also describes several common types of ANNs, focusing on perceptrons and multi-layer perceptrons.
Artificial neural networks (ANNs) are computing systems inspired by biological neural networks. ANNs can learn complex patterns and make predictions based on large amounts of data. The document discusses the basic structure and functioning of ANNs, including their ability to learn through adjustment of synaptic weights between neurons. It also describes several common types of ANNs, focusing on perceptrons and multi-layer perceptrons.
Pattern Recognition using Artificial Neural NetworkEditor IJCATR
An artificial neural network (ANN) usually called neural network. It can be considered as a resemblance to a paradigm
which is inspired by biological nervous system. In network the signals are transmitted by the means of connections links. The links
possess an associated way which is multiplied along with the incoming signal. The output signal is obtained by applying activation to
the net input NN are one of the most exciting and challenging research areas. As ANN mature into commercial systems, they are likely
to be implemented in hardware. Their fault tolerance and reliability are therefore vital to the functioning of the system in which they
are embedded. The pattern recognition system is implemented with Back propagation network and Hopfield network to remove the
distortion from the input. The Hopfield network has high fault tolerance which supports this system to get the accurate output.
Deep learning is a collection of machine learning algorithms utilizing multiple layers, with which higher levels of raw data are slowly removed. For example, lower layers can recognize edges in image processing whereas higher layers may define concepts for humans such as numbers or letters or faces. In this paper we have done a literature survey of some other papers to know how useful is Deep Learning and how to define other Artificial Intelligence things using Deep Learning. Anirban Chakraborty "A Study of Deep Learning Applications" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-4 | Issue-4 , June 2020, URL: https://www.ijtsrd.com/papers/ijtsrd31629.pdf Paper Url :https://www.ijtsrd.com/computer-science/artificial-intelligence/31629/a-study-of-deep-learning-applications/anirban-chakraborty
Similar to 88b984bd3f229c56fc7f4597d4e785f2 (1).pdf (20)
This document provides an overview of medical image segmentation using deep learning techniques. It discusses several deep learning architectures used for medical image segmentation, including U-Net, V-Net, GoogleNet, and ResNet. U-Net uses a symmetric encoder-decoder structure with skip connections to efficiently segment biomedical images. V-Net directly processes 3D MRI volumes for prostate segmentation. GoogleNet and ResNet employ inception modules and residual connections, respectively, to reduce parameters and enable training of very deep networks for medical image analysis tasks. The document aims to classify medical image segmentation approaches, discuss challenges, and outline future research directions using deep learning.
This document presents a comparative study of brain tumor segmentation and detection techniques in MRI images. It discusses several techniques used for brain tumor segmentation including fuzzy transform, morphological operations, thresholding, edge-based and region-based methods. The document also reviews various literature on brain tumor detection algorithms using techniques such as fuzzy clustering, neural networks, watershed segmentation, and support vector machines. Accuracy levels of different algorithms are presented ranging from dice coefficients above 0.90 to overall accuracy of 96.51%.
This document summarizes an article on using genetic algorithms for feature selection on brain tumor datasets. It discusses different feature selection methods like filter, wrapper and embedded methods. Specifically, it covers forward selection, backward elimination, recursive feature elimination, and genetic algorithms. It then reviews literature applying these various feature selection techniques to brain tumor classification problems. The goal is to identify the most important features to improve the accuracy of brain tumor detection systems.
Convolutional neural networks (CNNs) are commonly used for image classification and recognition tasks. CNNs use convolutional layers that apply learnable filters to detect patterns in input images. The filters are convolved across the width and height of the input to produce an activation map. Padding is added to images processed by CNNs to allow filters to fully cover edge pixels and produce accurate analyses. The dimensions of convolved outputs depend on the input and filter sizes, as well as any padding used.
This document discusses a study that proposes a framework for classifying brain tumors using an ensemble of deep features extracted from pre-trained convolutional neural networks (CNNs) and machine learning (ML) classifiers. The framework uses 13 pre-trained CNNs to extract deep features from magnetic resonance (MRI) brain images, which are then evaluated and the top 3 features selected using 9 ML classifiers. The selected features are concatenated to create an ensemble feature, which is classified using ML classifiers. The study evaluates this approach on 3 brain MRI datasets with different numbers of classes to classify tumors. Experimental results show that ensembling deep features can improve performance significantly, and support vector machines generally perform best, especially on larger datasets.
Up growth an efficient algorithm for high utility itemset mining(sigkdd2010) (1)vinayaga moorthy
The document proposes the UP-Growth algorithm for efficiently mining high utility itemsets. It constructs a UP-Tree structure to maintain high utility itemset information. It uses four strategies - discarding global unpromising items, discarding global node utilities, discarding local unpromising items, and decreasing local node utilities - to reduce the search space and number of candidate itemsets, improving mining performance. Experimental results show the algorithm scales well.
8 Surprising Reasons To Meditate 40 Minutes A Day That Can Change Your Life.pptxHolistified Wellness
We’re talking about Vedic Meditation, a form of meditation that has been around for at least 5,000 years. Back then, the people who lived in the Indus Valley, now known as India and Pakistan, practised meditation as a fundamental part of daily life. This knowledge that has given us yoga and Ayurveda, was known as Veda, hence the name Vedic. And though there are some written records, the practice has been passed down verbally from generation to generation.
These lecture slides, by Dr Sidra Arshad, offer a quick overview of the physiological basis of a normal electrocardiogram.
Learning objectives:
1. Define an electrocardiogram (ECG) and electrocardiography
2. Describe how dipoles generated by the heart produce the waveforms of the ECG
3. Describe the components of a normal electrocardiogram of a typical bipolar lead (limb II)
4. Differentiate between intervals and segments
5. Enlist some common indications for obtaining an ECG
6. Describe the flow of current around the heart during the cardiac cycle
7. Discuss the placement and polarity of the leads of electrocardiograph
8. Describe the normal electrocardiograms recorded from the limb leads and explain the physiological basis of the different records that are obtained
9. Define mean electrical vector (axis) of the heart and give the normal range
10. Define the mean QRS vector
11. Describe the axes of leads (hexagonal reference system)
12. Comprehend the vectorial analysis of the normal ECG
13. Determine the mean electrical axis of the ventricular QRS and appreciate the mean axis deviation
14. Explain the concepts of current of injury, J point, and their significance
Study Resources:
1. Chapter 11, Guyton and Hall Textbook of Medical Physiology, 14th edition
2. Chapter 9, Human Physiology - From Cells to Systems, Lauralee Sherwood, 9th edition
3. Chapter 29, Ganong’s Review of Medical Physiology, 26th edition
4. Electrocardiogram, StatPearls - https://www.ncbi.nlm.nih.gov/books/NBK549803/
5. ECG in Medical Practice by ABM Abdullah, 4th edition
6. Chapter 3, Cardiology Explained, https://www.ncbi.nlm.nih.gov/books/NBK2214/
7. ECG Basics, http://www.nataliescasebook.com/tag/e-c-g-basics
Adhd Medication Shortage Uk - trinexpharmacy.comreignlana06
The UK is currently facing a Adhd Medication Shortage Uk, which has left many patients and their families grappling with uncertainty and frustration. ADHD, or Attention Deficit Hyperactivity Disorder, is a chronic condition that requires consistent medication to manage effectively. This shortage has highlighted the critical role these medications play in the daily lives of those affected by ADHD. Contact : +1 (747) 209 – 3649 E-mail : sales@trinexpharmacy.com
Histololgy of Female Reproductive System.pptxAyeshaZaid1
Dive into an in-depth exploration of the histological structure of female reproductive system with this comprehensive lecture. Presented by Dr. Ayesha Irfan, Assistant Professor of Anatomy, this presentation covers the Gross anatomy and functional histology of the female reproductive organs. Ideal for students, educators, and anyone interested in medical science, this lecture provides clear explanations, detailed diagrams, and valuable insights into female reproductive system. Enhance your knowledge and understanding of this essential aspect of human biology.
- Video recording of this lecture in English language: https://youtu.be/kqbnxVAZs-0
- Video recording of this lecture in Arabic language: https://youtu.be/SINlygW1Mpc
- Link to download the book free: https://nephrotube.blogspot.com/p/nephrotube-nephrology-books.html
- Link to NephroTube website: www.NephroTube.com
- Link to NephroTube social media accounts: https://nephrotube.blogspot.com/p/join-nephrotube-on-social-media.html
TEST BANK For An Introduction to Brain and Behavior, 7th Edition by Bryan Kol...rightmanforbloodline
TEST BANK For An Introduction to Brain and Behavior, 7th Edition by Bryan Kolb, Ian Q. Whishaw, Verified Chapters 1 - 16, Complete Newest Versio
TEST BANK For An Introduction to Brain and Behavior, 7th Edition by Bryan Kolb, Ian Q. Whishaw, Verified Chapters 1 - 16, Complete Newest Version
TEST BANK For An Introduction to Brain and Behavior, 7th Edition by Bryan Kolb, Ian Q. Whishaw, Verified Chapters 1 - 16, Complete Newest Version
1. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2359
NEURO-FUZZY BASED MRESNET (NFMRESNET)
CLASSIFICATION FOR BRAIN TUMOR IMAGE
DATASET
R. Vinayaga Moorthy1
, R. Balasubramanian2
1
Research scholar, Department of Computer Science & Engineering, Manonmaniam Sundaranar University,
Tirunelveli, Tamil Nadu, India-627 012
2
Professor, Department of Computer Science & Engineering, Manonmaniam Sundaranar University, Tirunelveli,
Tamil Nadu, India-627 012
Email: vinayagamoorthy3@gmail.com1
,rbalus662002@yahoo.com
2
Article History: Received: 16.04.2023 Revised: 03.05.2023 Accepted: 19.06.2023
Abstract: Brains are enormous and complex organs that control our nervous systems and contain about 100 billion
nerve cells. The brain is an essential organ. A brain abnormality could put human health at risk. Tumors in the
brain are among the most serious of these abnormalities. An uncontrollable growth of brain cells inside the skull
causes this serious form of cancer. Generally, tumor cells exhibit heterogeneity, making them difficult to classify. In
order to decide on the correct medication, it is essential that tumors are detected early, and their location, size, and
types must also be assessed. Developing systems that incorporate human expertise is becoming increasingly popular
using Soft Computing. Image processing and cytology are used more often to diagnose disease. Correct diagnosis is
essential in treating and curing diseases. This paper proposes a fuzzy logic based brain tumor classification method
that can be used for proper treatment planning. This paper provides detailed analysis of the advantages of the
hybrid method, demonstrating the fact that when Neuro-Fuzzy Neural is paired with MResNet (NFMResnet), there is
a significant increase in classification accuracy. The NFMResnet contains convolutional layers, pooling layers, and
fully-connected layers, as well as a Fuzzy Self-Organization Layer. Using MResNet and fuzzy logic, the model
handles uncertain and imprecise input patterns. Three independent steps are involved in training the NFMResnet.
Three independent steps are involved in training the NFMResnet.
Keywords: Fuzzy C-Means Clustering, MResnet, Genbourta, NFMResnet
1. INTRODUCTION
Tumors are swellings that form lumps or masses in the body. A tumor is a lump or mass produced by a
pathological process within the body. It can be referred to as a swelling. A neoplasm is characterized by
tumors. Cancers are usually referred to as neoplasms. In some cases, doctors may confuse infections with
tumors when performing image diagnosis. In some cases, body cells may lose their ability to respond to
physiological signals. Such tissues are controlled by physiological mechanisms. This results in tumors
getting into place. Tumors are formed from uncontrolled growth of body cells and are referred to as
neoplastic tissue. Among the structures in the brain connected to tumors are neurons, blood vessels,
skulls, lymphatic tissues, pituitaries, and pineal glands.
2. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2360
Fig 1: Classification Process
Although images of complex real-world scenes or complex objects are difficult to classify, it is still
difficult to detect the boundaries between classes. There is often uncertainty in the representation of these
classification objects, and they have complex structures with overlapping, non isolated classes.
The use of MResNet in image classification is one of the most powerful approaches today. In MResNet
architectures, inputs are explicitly assumed to be images, allowing certain abstract properties to be
encoded.
Fuzzy classification, unlike classical classification, has continuous boundaries with overlapping areas
between neighboring classes. The degree to which an object belongs to different classes determines its
classification. This is a useful approach for a wide variety of applications, as well as for representing
complex feature spaces in a simple way.
1.1 Feature Extraction
It is easier to comprehend an image if we extract its most significant features. Feature extraction is the
process of identifying features. An image texture feature is extracted using GLCMs (grey-level co-
occurrence matrix). In addition to enhancing the details of the image, it gives interpretations to it. In
computer vision and image processing, shape descriptors are a powerful tool used to match objects,
classify them, recognize and identify them. Histogram-based detection is more specific when combined
with other detection methods such as shape analysis and edge finder based on Fourier transforms.
Training Image
Glioma Tumor 826 Images
Meningioma Tumor 822 Images
Pituitary Tumor 827 Images
No Tumor 395 Images
Testing Image
Glioma Tumor 100 Images
Meningioma Tumor 115 Images
Pituitary Tumor 74 Images
No Tumor 105 Images
Feature Extraction
Feature Selection
Conv Pooling Conv Pooling Conv Fuzzy Layer
MResNet Classifier
3. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2361
1.2 Feature Selection
The Boruta Algorithm and the Genetic Algorithm are combined in a hybrid algorithm. In terms of
selecting optimal feature subsets from a limited set of features, Genboruta uses the advantages of existing
algorithms. In order to select the most appropriate features from the data extracted from feature
extraction, the feature selection method is applied. The most significant features are selected using the
Genboruta algorithm.
1.3 Fuzzy Based Classification
Researchers have been paying a lot of attention to Deep Learning, the newest and most popular trend in
the machine learning field. Throughout various fields, deep learning has been used as a powerful machine
learning tool to solve complex problems, particularly those requiring highly accurate and sensitive results.
Comparatively to classical classifications, fuzzy classifications have overlapping areas between
neighboring classes. The degree to which an object belongs to different classes determines its
classification. A simple representation of a complex feature space is provided by this approach, which can
be used for a wide variety of applications.
Fuzzy logic can be incorporated into neural networks to make them more adaptable to cognitive
uncertainties. A fuzzy neural network is a hybrid network whose outputs are called neural fuzzy, neuro-
fuzzy or fuzzy-neuro networks. Fuzzy computations and neural networks are combined in hybrid systems
Practically, fuzzy neural networks are more effective than fuzzy neural networks or ordinary (classical)
neural networks, because fuzzy neural networks allow indeterminate and inaccurate information
processing. The transition from one cluster to another can be more gradual with fuzzy clustering as it is
more robust to outliers and noise in the data.
1.4 Fuzzy C-Means Clustering
It is possible with fuzzy clustering to group data points into more than one cluster with varying degrees of
membership. The fuzzy clustering algorithm assigns membership degrees between 0 and 1 to each data
point rather than assigning them to a single cluster like k-means or hierarchical clustering. If the data has
a complex structure or overlapping class boundaries, fuzzy clustering can be useful. In addition to
providing a more accurate understanding of the data structure, fuzzy clustering also allows for more
detailed representation of relationships between data points and clusters.
2 LITERATURE SURVEY
For many years, artificial intelligence has struggled to solve problems that were deemed insoluble by deep
learning. In addition to speech recognition [17], natural language processing [11], information retrieval
[6], computer vision [5], biomedicine [5], and social media analysis [1], it is exceptionally adept at
discovering intricate structures in high dimensional data.
Natural data in their raw form cannot be processed by conventional machine-learning techniques.
Machine learning or pattern recognition systems have been constructed for decades with careful
engineering and domain expertise. In order for a learning system (often a classifier) to detect or classify
patterns in input data, the data has to be transformed into a suitable internal representation or feature
vector. Deep learning techniques, which differ from conventional neural networks, use multiple hidden
layers to determine representations needed for classification or detection based on raw data [20]. The
methods used in deep learning include feedforward deep neural networks (DNNs)[7], convolutional
neural networks (CNNs) [12], recurrent neural networks (RNNs) [16], spiked neural networks (SNNs)
[9], long short-term memory recurrent networks (LSTMRNs) [18], stacked auto-encoders (SAEs)[21],
deep belief networks (DBNs)[23], and restricted Boltzmann machines (RBMs)[13].
An adaptable neuro-fuzzy inference system (ANFIS) was recently proposed as a model for addressing this
issue, and several successful attempts have been reported in the literature [2–4]. To form deep neurofuzzy
systems (DNFS), these studies combined deep neural networks with fuzzy logic (FL). Through the use of
fuzzy IF–THEN rules, this hybridization of DNN and FL effectively reduced uncertainty. It has been
4. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2362
rapidly growing in popularity since DNFS was introduced to solve a variety of real-world problems,
including Ramasamy and Hameed [3] proposed fuzzy convolutional neural network (FCNN) using both
FL and DNN to classify healthcare data into categories. There are two main parts of FCNN training:
parameter initialization and fine-tuning. A FCNN classifier was used in this study to classify ambiguous
or noisy data. According to the results of the study, the proposed FCNN method can easily remove
uncertainties and noise from original data.
As Price et al. reported [22], using best-in-class pre-prepared models, AlexNet, VGG16, GoogleLeNet,
Inception-v3, and ResNet-18, the fuzzy layers can be used for deep learning, enabling a wide selection of
combinations and yields.
Another classifier reliant upon fuzzy logic and wavelet change in a brain network was depicted in this
review. A layer in this classifier predicts the mathematical trademark related with marks or
characterizations. The proposed classifier is utilized to analyze cerebrum growths [10].
Among the most exploited techniques in image processing is CNN. Clinical diagnosis has assumed an
increasing significance in contemporary healthcare systems because it is capable of recognizing patterns
in images.U-NET Convolutional Neural Network (CNN) classification methodology and edge detection
fuzzy logic are used in this article in order to develop a method for the detection of brain tumors using
edge detection and fuzzy logic [19].
3 METHODOLOGY
Anatomical and functional information about brain tumors contributes greatly to enhancing diagnoses and
simplifying disease treatment planning in medical imaging applications. A brain tumor analysis can,
however, be impacted by the presence of image artifacts such as noise, intensity in homogeneity, and
partial volume effects. It is also necessary to consider the complex anatomy of the brain. Visual content
is extracted from images for indexing and retrieval through feature extraction. Basic features of an image
may be general in nature, such as color, texture, and shape, or they may be domain-specific. Features are
extracted by reducing the number and creating a new set of them that have the same information as the
originals, but that are completely different. These methods improve the accuracy of the classifier,
minimize overfitting, allow visualization of data, and increase training speed.
Machine learning algorithms are applied to a subset of features in the data by feature selection (also
known as subset selection). The best subset has the fewest dimensions that contribute to high accuracy;
the rest are discarded.
Recent years have seen DL's performance improve in several domains. In DL models, multiple levels of
information can be learned automatically from a large set of data. Their advantage is that they do not
require a lot of effort for tuning the features and expert knowledge like traditional machine learning. DL
has several architecture. Image processing uses CNN as a technique for recognizing patterns in images.
5. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2363
Fig 2: Proposed Fuzzy Based MResnet Classification
3.1 Feature Extraction
By analyzing images and objects, feature extraction methodologies extract features that represent various
classes of objects. A classifier assigns features to their respective classes based on their features. A feature
extraction algorithm reduces original data by identifying certain properties that separate one input pattern
from another. In the extracted feature vector, the relevant properties of the image should be described in
order to provide the classifier with the characteristics of the input type.
3.2 Features Optimization Using GenBoruta
This research relies on using a hybrid model that combines both a Genetic Algorithm and a Boruta
Algorithm, with an original aim of grouping tests before selecting the small number of important
variables.The use of a hybrid technique overcomes the disadvantages of each individual one [14].
Table 1: Selected Features
S.No Genboruta (Features)
1. Surface Area 2. Range
3. Flatness 4. Cluster shade
5. Skewness 6. Dissimilarity
7. Uniformity 8. Maximum Probability
9. Contrast 10. Variance
11. Correlation 12. Max Intensity
6. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2364
3.3 Modified Resnet (MResnet)
RESNET18 architecture consists of 18 layers. There are three layers of convolution, each using three
kernels. Stack layers are added to the output; therefore, complex computations are not required with the
MResNet algorithm because one or more layers are skipped. A shortcut connection for ResNet18 omits
two layers. The design was based on Swish activation functions [15].
Fig 3: MResNet Classification Techniques
3.4. The Structure of an NFMResnet
The proposed MRNN's architecture features four types of layers: convolutional layer, pooling layer, self-
organization (or fuzzy) layer, and fully connected layer. MResnet Fuzzy Neural Networks are built by
stacking three parts:
A MResnet network (convolutional and pooling Layers);
The Self-Organization Layer (The Fuzzy Layer);
A classifier (some fully-connected layers).
Fig 4: Struture of NFMResnet
To classify real-world objects and scene images, we propose an MResnet fuzzy neural network
(NFMResnet). A self-organization layer is utilized to provide preprocessing in the proposed NFMResnet
model; unlike a regular MResnet. To cope with uncertainties and ambiguities in input data, the proposed
network incorporates the advantages of MResnet and FL Three layers of the NFMResnet are shown in
Figure 4: the convolutional network, the self-organization layer, and the classifier layer. Below is a
description of each part's functionality:
7. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2365
1 The first layer of the network is the convolutional network, which substitutes the pooling layer with
the convolutional layer for abstraction at the high level.
2 Secondly, the fuzzy layer divides the input data into predetermined clusters before making a final
clustering decision. Fuzzy layer output neurons are a representation of the fuzzy input cluster
membership functions, with membership grades reflecting the degree to which data points are related
to each cluster.
3 A classifier is used to calculate the class score as an output of the NFMResnet network, the third part
of the network.
Fuzzy C-Means Clustering
Data clusters are detectable with the FCM algorithm, but these clusters do not have any order, making
interpretation challenging. In this paper, the cluster prototypes are ordered on an easily visualizable small
dimension space to increase the transparency of the clustering result. A regularization of the clusters can
achieve this ordering. The original objective function of FCM is penalized in order to achieve such
regularization. On a small dimensional space, cluster centers are laid out on a grid for easy visualization.
Regularizing the fuzzy c-means function (FCM) results in the smoothness of this mapping.
NFMResnet uses a radial basis function to model the membership of an input vector x to each of the L
clusters, whereas the number of neurons in the fuzzy layer is L.
𝑓(𝑥) =
√
e
( )
1
In the above condition, m and σ - are two genuine qualities, where the focal point of a bunch is addressed
as m, and σ - is utilized to obscure the limit level of a group. A compound interest expression arising in
the study of compound interest is the Euler's number 'e'. The sum of infinite numbers can also be
expressed in this way.
0
!
n
n
e
𝜇𝑙 x)( )
= 𝑓(𝑠) = 𝑓(𝑓 𝑥 ( ) 2
𝜇𝑙 x)( )
= 1 3
The hybrid network formed a vector consisting of degrees of belonging to the specific cluster centers
based on the vector x = [x1, x2, ..., xn] fed to its input: [μ1(x),μ2(x), ..., μL(x)]. In the "fuzzy layer", the
outputs of neurons are used as inputs to the classifier, which has been conditioned to satisfy the
normalization condition (3) for each training sample vector x(k). An input to the classifier is the output of
neurons in the "fuzzy layer".
NFMResnet works in three stages: first, it transforms an image into a vector of high-level characteristics,
second, the fuzzy layer distributes the input data into fuzzy clusters; thirdly, the final fully connected
layer performs the classification by assigning a label to each group of clusters based on the result class.
3.5 The Training of the NFMResnet
MResnet's training process involves three independent steps for each component of the neural network.
The MResnet network is first trained by backpropagation to form some abstract properties of the input
image .In today's world, most "pre-trained" models have already been trained on a large amount of data
from a related domain.
It is the process of tuning the fuzzy layer parameters, called self-organization,that is the second part of the
process. Fuzzy Layers have a self-organizing nature. An unsupervised competitive learning scheme is
used to train it. In "fuzzy layers", centers of clusters are self-organized.
8. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2366
Training the classifier is the third part. The MResnet and fuzzy layers have stable parameters. Tuning is
only done on the weights of fully connected layers. A standard backpropagation algorithm is used to train
the classifier.
A pixel array of an image is fed into the NFMResnet input after NFMResnet has completed all three parts
of training. y = [y1, y2, ..., yp] represents a vector that represents whether a given image belongs to each
class (class scores) in the input image. Images are assigned to classes based on their maximum score.
Fig 5: Architecture of Proposed NFMResnet Classification Technique
NFMResnet Algorithm
9. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2367
4. RESULTS
CFNN is used in some experiments. An image is classified as either having a tumor or not using the
model. The dataset used here is Kaggle's ' Brain tumor '. The training set contains 2,870 MRI brain tumor
images, while the test set contains 394.
In Figure 6, we show some samples of images from the 'Brain Tumor' dataset.
Fig 6: Sample MRI Brain Tumor Image
The first step is to use a pretrained model that recognizes a wide range of images and then fine-tune it for
binary classification. In order to achieve a simple yet powerful neural network architecture, we chose
Lenet, AlexNet, VGG Net, and Resnet models pretrained on the Kaggle dataset. To train the NFMResnet
model, we took three independent steps:
1. Analyzing brain tumor images to classify them using Lenet, AlexNet, VGG Net, and Resnet models
(stochastic optimization algorithm: Adam). Resources are required for the MResnet training.
2. Clustering based on fuzzy c-means (self-organizing fuzzy layer). Multiple clusters have been created
with the data set, each with a different number of clusters. Using the Fuzzy Partition Coefficient, we
choose the number of clusters when it is maximized (FPC varies from 0 to 1, with 1 being the best. It is a
metric that measures how well the clustering model describes the data).
3. A stochastic optimization method called Adam is used for classifier training. The parameters of the
MResnet and fuzzy layers were stable while only the weights of the fully-connected layers were being
tuned.
The results of experiments show that incorporating the fuzzy layer into MResnet improve classification
problem quality (accuracy) despite the regular CNN not showing exceptional accuracy.
4.1 Evaluation Metrics
10. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2368
Table 6: Classification Performance Analysis
Fig 3: Classification Graphical Representation
5. CONCLUSION
A brief discussion of all relevant data must precede the conclusion of this extensive research. We
compared its performance against pretrained LeNet, AlexNet, VGGNet, and ResNet, MResnet for the
brain tumor analysis. Three thousand two hundred sixty-four records make up the Kaggle brain tumor
data set. A stratified sampling method was used to divide the data set into 10 mutually exclusive
partitions for model building and evaluation. The training partitions were used for seven, while the testing
partitions were used for four. For each of the four models, accuracy, precision, recall, and F1Score are
calculated.
Therefore, this paper presents an image classification model based on Neuro fuzzy MResnet
(NFMResnet). In terms of fuzzy C- Maens Cluster, fuzzyness is incorporated into the structure of the
network. MResnet is combined with fuzzy logic in the proposed model for handling uncertainty and
imprecision. Experiments measuring the effectiveness of NFMResnet have been conducted and indicate
that the NFMResnet could provide better accuracy.
0
10
20
30
40
50
60
70
80
90
100
Accuracy
Precision
Recall
Specificity
F1-Score
11. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2369
REFERENCES
1. Andrei Romascanu, Hannah Ker, Renee SieberSarah Greenidge, Sam Lumley, Drew Bush, Stefan
Morgan, Rosie Zhao and Mikael Brunila, “Using deep learning and social network analysis to
understand and manage extreme flooding” , John Wiley & Sons, pp.251-261,2020.
2. Angelica I. Aviles, Samar M. Alsaleh, Eduard Montseny, Pilar Sobrevilla & and Alicia Casals, “A
Deep-Neuro-Fuzzy Approach for Estimating theInteraction Forces in Robotic Surgery”, IEEE,
pp.1113-1119,2016.
3. Balamurugan Ramasamy, Abdul Zubar Hameed, “Classification of healthcare data using hybridised
fuzzy and convolutional neural network”, Healthcare Technology Letters, Volume 6, Issue No. 3, pp.
59–63, 2019.
4. Chaimae El Hatri, Jaouad Boumhidi, “Fuzzy deep learning based urban traffic incident detection”,
ScienceDirect, Elsevier, pp.1-8,2017.
5. Eralda Nishani, Betim Cico, “Computer Vision Approaches based on Deep Learning and Neural
Networks: Deep Neural Networks for Video Analysis of Human Pose Estimation”, IEEE, PP. 11-15,
2017.
6. Jiafeng Guo, Yixing Fan, Liang Pang, Liu Yang, Qingyao Ai, Hamed Zamani, Chen Wu, W. Bruce
Croft, Xueqi Cheng, “A Deep Look into neural ranking models for information retrieval’ ,
Information Processing and Management, Elsevier Ltd, pp.1-20,2019.
7. Maximilian Schaedler, Christian Bluemm, Maxim Kuschnerov, Fabio Pittala, Stefano Calabro and
Stephan Pachnicke, “Deep Neural Network Equalization for Optical Short Reach Communication” ,
Applied Science, MDPI, Volume 9, pp.1-14,2019.
8. Michael Wainberg, Daniele Merico, Andrew Delong & Brendan J Frey, “Deep learning in
biomedicine”, nature biotechnology, Volume 36 ISSUE Number 9,PP.829-838,2018.
9. Mohd Hafizul Afifi Abdullah, Muhaini Othman, Shahreen Kasim, Siti Aisyah Mohamed, “Evolving
spiking neural networks methods for classification problem: a case study in flood events risk
assessment”, Indonesian Journal of Electrical Engineering and Computer Science, Vol. 16, Issue No.
1, pp. 222-229, 2019.
10. Mohsen Ahmadi ,Fatemeh Dashti Ahangar,Nikoo Astaraki,Mohammad Abbasi ,and Behzad Babaei,
“FWNNet: Presentation of a New Classifier of Brain Tumor Diagnosis Based on Fuzzy Logic and the
Wavelet-Based Neural Network Using Machine-Learning Methods”, Computational Intelligence and
Neuroscience, Hindawi, pp1-13,2021.
11. N I Widiastuti, “Convolution Neural Network for Text Mining and Natural Language Processing”
,Materials Science and Engineering, IOP Publishing, pp.1-6,2019.
12. Najwa Altwaijry and Isra Al-Turaiki, “Arabic handwriting recognition system using convolutional
neural network”, Neural Computing and Applications, Springer, pp.1-13, 2020.
13. Nan Zhang , Shifei Ding , Jian Zhang , Yu Xue, “An overview on restricted boltzmann machines”,
Neurocomputing, pp.1-30, 2017.
14. R.Vinayaga Moorthy, Dr.R.Balasubramanian, “An Efficient Genetic Boruta(GenBoruta) Algorithm
Based Feature Selection on Brain Tumor Dataset” , International Journal of Mechanical Engineering,
Vol. 7,Issue No. 6,pp.694-712, 2022.
15. R.Vinayagamoorthy, Dr.R.Balasubramanian, “A Modified Resnet (MRestNet) based Classification
for Brain Tumor Dataset”, NeuroQuantology, Volume 20, Issue 15,Page 7750-7767, 2022.
16. Robin M. Schmidt, “Recurrent Neural Networks (RNNs):A gentle Introduction and Overview” ,
arXiv, pp. 1-16, 2019.
17. S. Sharmadha, K. Shivani, K. Shruthi, B. Bharathi and S. Kavitha, “Automatic Speech Recognition
Using Deep Neural Network” , Advances in Intelligent Systems and Computing, Springer, Volume
1118, pp. 353-361,2019.
18. Said Jadid Abdulkadir, Hitham Alhussian, Muhammad Nazmi, Asim A Elsheikh, “Long Short Term
Memory Recurrent Network for Standard and Poor’s 500 Index Modeling” , International Journal of
Engineering & Technology, Volume 7,Issue No 4.15,pp. 25-29, 2018.
12. Neuro-Fuzzy Based Mresnet (Nfmresnet) Classification For Section A-Research paper
Brain Tumor Image Dataset
Eur. Chem. Bull. 2023, 12 (6), 2359 – 2370 2370
19. Sarmad Maqsood, Robertas Damasevicius, Faisal Mehmood Shah, “An Efficient Approach for the
Detection of Brain Tumor Using Fuzzy Logic and U-NET CNN Classification” , Computational
Science and Its Applications, Page.105–118, 2021.
20. Yann LeCun, Yoshua Bengio & Geoffrey Hinton, “Deep learning”, Nature, Volume 521, pp.436-444,
2015.
21. Yi Qiu, Yiyang Dai, “A Stacked Auto-Encoder Based Fault Diagnosis Model for Chemical Process”,
Eindhoven, Elsevier, pp.1303-1308, 2019.
22. Yu-Jun Zheng, Member, Wei-Guo Sheng, Xing-Ming Sun, and Sheng-Yong Chen, “Airline
Passenger Profiling Based on Fuzzy Deep Machine Learning” , IEEE transactions on neural networks
and learning systems,pp.1-13,2016.
23. Yuming Hua, Junhai Guo, Hua Zhao, “Deep Belief Networks and Deep Learning” , IEEE,pp.1-4
,2015.