SlideShare a Scribd company logo
Content Based Image
Retrieval (CBIR)
Behzad Shomali
What is CBIR?
Content-based image retrieval, also known as query by image content (QBIC) and content-based
visual information retrieval (CBVIR), is the application of computer vision techniques to the image
retrieval problem, that is, the problem of searching for digital images in large databases.
https://en.wikipedia.org/wiki/Content-based_image_retrieval
Query
Image
Image
Feature
Extraction
Feature
Extraction
Similarity
Measuremen
t
Retrieved
Images
Technologies
● Query by example (QBE)


● Semantic retrieval


● Relevance feedback (human interaction)


● Iterative/machine learning


● Other query methods
https://en.wikipedia.org/wiki/Content-based_image_retrieval
Technologies
● Query by example (QBE)


● Semantic retrieval


● Relevance feedback (human interaction)


● Iterative/machine learning


● Other query methods
https://en.wikipedia.org/wiki/Content-based_image_retrieval
Application in popular search systems
● Google images


○ Constructing a mathematical model


○ Metadata


● eBay


○ ResNet-50 for category recognition


● SK Planet


○ inception-v3 as vision encoder


○ RNN multi-class classification


● Alibaba


○ GoogLeNet V1 for category prediction and feature learning


● Pinterest


○ Two-step object detection
https://en.wikipedia.org/wiki/Reverse_image_search
Application in popular search systems
● Google images


○ Constructing a mathematical model


○ Metadata


● eBay


○ ResNet-50 for category recognition


● SK Planet


○ inception-v3 as vision encoder


○ RNN multi-class classification


● Alibaba


○ GoogLeNet V1 for category prediction and feature learning


● Pinterest


○ Two-step object detection
https://en.wikipedia.org/wiki/Reverse_image_search
Image Representation and Features
● Extract local and deep features


● Studied AlexNet and VGG


○ Extract feature representations from fc6 and fc8 layers


○ Binarized


○ Hamming distance


● Extract salient color signatures


○ Detect salient regions


○ K-means clustering


○ Store cluster centroids and weights as image signature
[DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
Two-step Object Detection and Localization
1. Category classification


2. Object detection
[DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
Car Flower Person
c1 ... cn f1 ... fm p1 ... pk
Input
Category classification
Object detection
Car Flower Person
c1 ... cn f1 ... fm p1 ... pk
Input
Category classification
Object detection
Car Flower Person
c1 ... cn f1 ... fm p1 ... pk
Input
Category classification
Object detection
Reduce
computational cost
Static Evaluation of Search Relevance
● Used dataset contains: 1.6 M unique images


○ Be assumed to be relevant, if two images share a label


● Computed precision@k based on several features


○ The fc6 layer activations from the generic AlexNet (pre-trained for ILSVRC)


○ The fc6 activations of an AlexNet model fine-tuned to recognize over 3,000 Pinterest
products categories


○ The loss3/classifier activations of a generic GoogLeNet


○ The fc6 activations of a generic VGG 16-layer model
[DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
Precision vs. Recall
[Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601]
Precision vs. Recall
[Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601]
Either value alone contains insufficient information


● We can always make recall 1, simply by retrieving all images


● Similarly, precision can be kept high by retrieving only a few images
● P (10) ; P (30) ; P (NR) - the precision after the first 10 ; 30 ; NR documents are retrieved


● Mean Average Precision - mean (non-interpolated) average precision .


● recall at .5 precision - recall at the rank where precision drops below .5.


● R (1000) - recall after 1000 documents are retrieved.


● Rank first relevant - The rank of the highest-ranked relevant document.
Precision vs. Recall
[Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601]
Either value alone contains insufficient information


● We can always make recall 1, simply by retrieving all images


● Similarly, precision can be kept high by retrieving only a few images
Precision and recall
should either be
used together
● P (10) ; P (30) ; P (NR) - the precision after the first 10 ; 30 ; NR documents are retrieved


● Mean Average Precision - mean (non-interpolated) average precision .


● recall at .5 precision - recall at the rank where precision drops below .5.


● R (1000) - recall after 1000 documents are retrieved.


● Rank first relevant - The rank of the highest-ranked relevant document.
Relevance of visual search
Table 1 shows p@5 and p@10 performance of these models, along with the average CPU-based
latency of our visual search service, which includes feature extraction for the query image as well as
retrieval.
[DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
Siamese networks
[Das, Arpita, et al. "Together we stand: Siamese networks for similar question retrieval." Proceedings of the 54th Annual Meeting of the Association for
Computational Linguistics (Volume 1: Long Papers). 2016]
Siamese networks
● Let, F(X) be the family of functions with set of parameters W. F(X) is assumed to be
differentiable with respect to W. Siamese network seeks a value of the parameter W such that
the symmetric similarity metric is small if X1 and X2 belong to the same category, and large if
they belong to different categories.
[Das, Arpita, et al. "Together we stand: Siamese networks for similar question retrieval." Proceedings of the 54th Annual Meeting of the Association for
Computational Linguistics (Volume 1: Long Papers). 2016]
Different loss functions for training a Siamese network
Two commonly used ones are


● Triplet loss


● Contrastive loss


The main idea of these loss functions is to
pull the samples of every class toward one
another and push the samples of different
classes away from each other
[Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural
Networks (IJCNN). IEEE, 2020]
Different loss functions - Triplet loss
The triplet loss uses anchor, neighbor, and distant. Let f(x) be the output (i.e., embedding) of the network
for the input x. The triplet loss tries to reduce the distance of anchor and neighbor embeddings and desires
to increase the distance of anchor and distant embeddings. As long as the distances of anchor-distant pairs
get larger than the distances of anchor-neighbor pairs by a margin α ≥ 0, the desired embedding is obtained
[Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural
Networks (IJCNN). IEEE, 2020]
Different loss functions - Contrastive loss
The contrastive loss uses pairs of samples which can be anchor and neighbor or anchor and distant. If the
samples are anchor and neighbor, they are pulled towards each other; otherwise, their distance is
increased. In other words, the contrastive loss performs like the triplet loss but one by one rather than
simultaneously. The desired embedding is obtained when the anchor-distant distances get larger than the
anchor-neighbor distances by a margin of α
[Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural
Networks (IJCNN). IEEE, 2020]

More Related Content

What's hot

Image segmentation in Digital Image Processing
Image segmentation in Digital Image ProcessingImage segmentation in Digital Image Processing
Image segmentation in Digital Image Processing
DHIVYADEVAKI
 
Object recognition
Object recognitionObject recognition
Object recognition
saniacorreya
 
Lec1: Medical Image Computing - Introduction
Lec1: Medical Image Computing - Introduction Lec1: Medical Image Computing - Introduction
Lec1: Medical Image Computing - Introduction
Ulaş Bağcı
 
Image processing SaltPepper Noise
Image processing SaltPepper NoiseImage processing SaltPepper Noise
Image processing SaltPepper Noise
Ankush Srivastava
 
Content based image retrieval
Content based image retrievalContent based image retrieval
Content based image retrieval
rubaiyat11
 
IMAGE SEGMENTATION TECHNIQUES
IMAGE SEGMENTATION TECHNIQUESIMAGE SEGMENTATION TECHNIQUES
IMAGE SEGMENTATION TECHNIQUES
Vicky Kumar
 
Removal of Salt and Pepper Noise in images
Removal of Salt and Pepper Noise in imagesRemoval of Salt and Pepper Noise in images
Removal of Salt and Pepper Noise in images
Murali Siva
 
Object Recognition
Object RecognitionObject Recognition
Object Recognition
Eman Abed AlWahhab
 
Threshold Selection for Image segmentation
Threshold Selection for Image segmentationThreshold Selection for Image segmentation
Threshold Selection for Image segmentation
Parijat Sinha
 
Active contour segmentation
Active contour segmentationActive contour segmentation
Active contour segmentation
Nishant Jain
 
Image seg using_thresholding
Image seg using_thresholdingImage seg using_thresholding
Image seg using_thresholding
Vani011
 
Computer vision
Computer vision Computer vision
Computer vision
Dmitry Ryabokon
 
IMAGE SEGMENTATION.
IMAGE SEGMENTATION.IMAGE SEGMENTATION.
IMAGE SEGMENTATION.
Tawose Olamide Timothy
 
Chap6 image restoration
Chap6 image restorationChap6 image restoration
Chap6 image restoration
ShardaSalunkhe1
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
ArtiKhanchandani
 
Computer Vision.pptx
Computer Vision.pptxComputer Vision.pptx
Computer Vision.pptx
GDSCIIITDHARWAD
 
Color model
Color modelColor model
Intro to deep learning
Intro to deep learning Intro to deep learning
Intro to deep learning
David Voyles
 
Thresholding.ppt
Thresholding.pptThresholding.ppt
Thresholding.ppt
shankar64
 
Object detection with deep learning
Object detection with deep learningObject detection with deep learning
Object detection with deep learning
Sushant Shrivastava
 

What's hot (20)

Image segmentation in Digital Image Processing
Image segmentation in Digital Image ProcessingImage segmentation in Digital Image Processing
Image segmentation in Digital Image Processing
 
Object recognition
Object recognitionObject recognition
Object recognition
 
Lec1: Medical Image Computing - Introduction
Lec1: Medical Image Computing - Introduction Lec1: Medical Image Computing - Introduction
Lec1: Medical Image Computing - Introduction
 
Image processing SaltPepper Noise
Image processing SaltPepper NoiseImage processing SaltPepper Noise
Image processing SaltPepper Noise
 
Content based image retrieval
Content based image retrievalContent based image retrieval
Content based image retrieval
 
IMAGE SEGMENTATION TECHNIQUES
IMAGE SEGMENTATION TECHNIQUESIMAGE SEGMENTATION TECHNIQUES
IMAGE SEGMENTATION TECHNIQUES
 
Removal of Salt and Pepper Noise in images
Removal of Salt and Pepper Noise in imagesRemoval of Salt and Pepper Noise in images
Removal of Salt and Pepper Noise in images
 
Object Recognition
Object RecognitionObject Recognition
Object Recognition
 
Threshold Selection for Image segmentation
Threshold Selection for Image segmentationThreshold Selection for Image segmentation
Threshold Selection for Image segmentation
 
Active contour segmentation
Active contour segmentationActive contour segmentation
Active contour segmentation
 
Image seg using_thresholding
Image seg using_thresholdingImage seg using_thresholding
Image seg using_thresholding
 
Computer vision
Computer vision Computer vision
Computer vision
 
IMAGE SEGMENTATION.
IMAGE SEGMENTATION.IMAGE SEGMENTATION.
IMAGE SEGMENTATION.
 
Chap6 image restoration
Chap6 image restorationChap6 image restoration
Chap6 image restoration
 
Computer Vision
Computer VisionComputer Vision
Computer Vision
 
Computer Vision.pptx
Computer Vision.pptxComputer Vision.pptx
Computer Vision.pptx
 
Color model
Color modelColor model
Color model
 
Intro to deep learning
Intro to deep learning Intro to deep learning
Intro to deep learning
 
Thresholding.ppt
Thresholding.pptThresholding.ppt
Thresholding.ppt
 
Object detection with deep learning
Object detection with deep learningObject detection with deep learning
Object detection with deep learning
 

Similar to Content Based Image Retrieval (CBIR)

Kernel based similarity estimation and real time tracking of moving
Kernel based similarity estimation and real time tracking of movingKernel based similarity estimation and real time tracking of moving
Kernel based similarity estimation and real time tracking of moving
IAEME Publication
 
Scene Description From Images To Sentences
Scene Description From Images To SentencesScene Description From Images To Sentences
Scene Description From Images To Sentences
IRJET Journal
 
2019 cvpr paper overview by Ho Seong Lee
2019 cvpr paper overview by Ho Seong Lee2019 cvpr paper overview by Ho Seong Lee
2019 cvpr paper overview by Ho Seong Lee
Moazzem Hossain
 
2019 cvpr paper_overview
2019 cvpr paper_overview2019 cvpr paper_overview
2019 cvpr paper_overview
LEE HOSEONG
 
Optimized Feedforward Network of CNN with Xnor Final Presentation
Optimized Feedforward Network of CNN with Xnor Final PresentationOptimized Feedforward Network of CNN with Xnor Final Presentation
Optimized Feedforward Network of CNN with Xnor Final Presentation
Indiana University Bloomington
 
物件偵測與辨識技術
物件偵測與辨識技術物件偵測與辨識技術
物件偵測與辨識技術
CHENHuiMei
 
Dataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problemsDataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problems
PetteriTeikariPhD
 
Cartoonization of images using machine Learning
Cartoonization of images using machine LearningCartoonization of images using machine Learning
Cartoonization of images using machine Learning
IRJET Journal
 
IRJET - Vehicle Classification with Time-Frequency Domain Features using ...
IRJET -  	  Vehicle Classification with Time-Frequency Domain Features using ...IRJET -  	  Vehicle Classification with Time-Frequency Domain Features using ...
IRJET - Vehicle Classification with Time-Frequency Domain Features using ...
IRJET Journal
 
AI and Deep Learning
AI and Deep Learning AI and Deep Learning
AI and Deep Learning
Subrat Panda, PhD
 
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
IRJET Journal
 
K-Means Clustering in Moving Objects Extraction with Selective Background
K-Means Clustering in Moving Objects Extraction with Selective BackgroundK-Means Clustering in Moving Objects Extraction with Selective Background
K-Means Clustering in Moving Objects Extraction with Selective Background
IJCSIS Research Publications
 
ArtificialIntelligenceInObjectDetection-Report.pdf
ArtificialIntelligenceInObjectDetection-Report.pdfArtificialIntelligenceInObjectDetection-Report.pdf
ArtificialIntelligenceInObjectDetection-Report.pdf
Abishek86232
 
Hand Written Digit Classification
Hand Written Digit ClassificationHand Written Digit Classification
Hand Written Digit Classification
ijtsrd
 
Image Generation from Caption
Image Generation from Caption Image Generation from Caption
Image Generation from Caption
IJSCAI Journal
 
IMAGE GENERATION FROM CAPTION
IMAGE GENERATION FROM CAPTIONIMAGE GENERATION FROM CAPTION
IMAGE GENERATION FROM CAPTION
ijscai
 
Introduction to Machine Learning with SciKit-Learn
Introduction to Machine Learning with SciKit-LearnIntroduction to Machine Learning with SciKit-Learn
Introduction to Machine Learning with SciKit-Learn
Benjamin Bengfort
 
Image classification with Deep Neural Networks
Image classification with Deep Neural NetworksImage classification with Deep Neural Networks
Image classification with Deep Neural Networks
Yogendra Tamang
 
最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に - 最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に -
Hiroshi Fukui
 
Learning with Relative Attributes
Learning with Relative AttributesLearning with Relative Attributes
Learning with Relative Attributes
Vikas Jain
 

Similar to Content Based Image Retrieval (CBIR) (20)

Kernel based similarity estimation and real time tracking of moving
Kernel based similarity estimation and real time tracking of movingKernel based similarity estimation and real time tracking of moving
Kernel based similarity estimation and real time tracking of moving
 
Scene Description From Images To Sentences
Scene Description From Images To SentencesScene Description From Images To Sentences
Scene Description From Images To Sentences
 
2019 cvpr paper overview by Ho Seong Lee
2019 cvpr paper overview by Ho Seong Lee2019 cvpr paper overview by Ho Seong Lee
2019 cvpr paper overview by Ho Seong Lee
 
2019 cvpr paper_overview
2019 cvpr paper_overview2019 cvpr paper_overview
2019 cvpr paper_overview
 
Optimized Feedforward Network of CNN with Xnor Final Presentation
Optimized Feedforward Network of CNN with Xnor Final PresentationOptimized Feedforward Network of CNN with Xnor Final Presentation
Optimized Feedforward Network of CNN with Xnor Final Presentation
 
物件偵測與辨識技術
物件偵測與辨識技術物件偵測與辨識技術
物件偵測與辨識技術
 
Dataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problemsDataset creation for Deep Learning-based Geometric Computer Vision problems
Dataset creation for Deep Learning-based Geometric Computer Vision problems
 
Cartoonization of images using machine Learning
Cartoonization of images using machine LearningCartoonization of images using machine Learning
Cartoonization of images using machine Learning
 
IRJET - Vehicle Classification with Time-Frequency Domain Features using ...
IRJET -  	  Vehicle Classification with Time-Frequency Domain Features using ...IRJET -  	  Vehicle Classification with Time-Frequency Domain Features using ...
IRJET - Vehicle Classification with Time-Frequency Domain Features using ...
 
AI and Deep Learning
AI and Deep Learning AI and Deep Learning
AI and Deep Learning
 
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
IRJET- Recognition of Handwritten Characters based on Deep Learning with Tens...
 
K-Means Clustering in Moving Objects Extraction with Selective Background
K-Means Clustering in Moving Objects Extraction with Selective BackgroundK-Means Clustering in Moving Objects Extraction with Selective Background
K-Means Clustering in Moving Objects Extraction with Selective Background
 
ArtificialIntelligenceInObjectDetection-Report.pdf
ArtificialIntelligenceInObjectDetection-Report.pdfArtificialIntelligenceInObjectDetection-Report.pdf
ArtificialIntelligenceInObjectDetection-Report.pdf
 
Hand Written Digit Classification
Hand Written Digit ClassificationHand Written Digit Classification
Hand Written Digit Classification
 
Image Generation from Caption
Image Generation from Caption Image Generation from Caption
Image Generation from Caption
 
IMAGE GENERATION FROM CAPTION
IMAGE GENERATION FROM CAPTIONIMAGE GENERATION FROM CAPTION
IMAGE GENERATION FROM CAPTION
 
Introduction to Machine Learning with SciKit-Learn
Introduction to Machine Learning with SciKit-LearnIntroduction to Machine Learning with SciKit-Learn
Introduction to Machine Learning with SciKit-Learn
 
Image classification with Deep Neural Networks
Image classification with Deep Neural NetworksImage classification with Deep Neural Networks
Image classification with Deep Neural Networks
 
最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に - 最近の研究情勢についていくために - Deep Learningを中心に -
最近の研究情勢についていくために - Deep Learningを中心に -
 
Learning with Relative Attributes
Learning with Relative AttributesLearning with Relative Attributes
Learning with Relative Attributes
 

Recently uploaded

Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
Dutch Power
 
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
Dutch Power
 
2024-05-30_meetup_devops_aix-marseille.pdf
2024-05-30_meetup_devops_aix-marseille.pdf2024-05-30_meetup_devops_aix-marseille.pdf
2024-05-30_meetup_devops_aix-marseille.pdf
Frederic Leger
 
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussion
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussionPro-competitive Industrial Policy – OECD – June 2024 OECD discussion
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussion
OECD Directorate for Financial and Enterprise Affairs
 
XP 2024 presentation: A New Look to Leadership
XP 2024 presentation: A New Look to LeadershipXP 2024 presentation: A New Look to Leadership
XP 2024 presentation: A New Look to Leadership
samililja
 
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussion
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussionPro-competitive Industrial Policy – LANE – June 2024 OECD discussion
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussion
OECD Directorate for Financial and Enterprise Affairs
 
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
OECD Directorate for Financial and Enterprise Affairs
 
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussionArtificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
OECD Directorate for Financial and Enterprise Affairs
 
ASONAM2023_presection_slide_track-recommendation.pdf
ASONAM2023_presection_slide_track-recommendation.pdfASONAM2023_presection_slide_track-recommendation.pdf
ASONAM2023_presection_slide_track-recommendation.pdf
ToshihiroIto4
 
Gregory Harris - Cycle 2 - Civics Presentation
Gregory Harris - Cycle 2 - Civics PresentationGregory Harris - Cycle 2 - Civics Presentation
Gregory Harris - Cycle 2 - Civics Presentation
gharris9
 
Burning Issue Presentation By Kenmaryon.pdf
Burning Issue Presentation By Kenmaryon.pdfBurning Issue Presentation By Kenmaryon.pdf
Burning Issue Presentation By Kenmaryon.pdf
kkirkland2
 
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie WellsCollapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
Rosie Wells
 
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
Suzanne Lagerweij
 
Competition and Regulation in Professions and Occupations – ROBSON – June 202...
Competition and Regulation in Professions and Occupations – ROBSON – June 202...Competition and Regulation in Professions and Occupations – ROBSON – June 202...
Competition and Regulation in Professions and Occupations – ROBSON – June 202...
OECD Directorate for Financial and Enterprise Affairs
 
Tom tresser burning issue.pptx My Burning issue
Tom tresser burning issue.pptx My Burning issueTom tresser burning issue.pptx My Burning issue
Tom tresser burning issue.pptx My Burning issue
amekonnen
 
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
gpww3sf4
 
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPointMẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
1990 Media
 
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
OECD Directorate for Financial and Enterprise Affairs
 
Gregory Harris' Civics Presentation.pptx
Gregory Harris' Civics Presentation.pptxGregory Harris' Civics Presentation.pptx
Gregory Harris' Civics Presentation.pptx
gharris9
 
Carrer goals.pptx and their importance in real life
Carrer goals.pptx  and their importance in real lifeCarrer goals.pptx  and their importance in real life
Carrer goals.pptx and their importance in real life
artemacademy2
 

Recently uploaded (20)

Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
Presentatie 4. Jochen Cremer - TU Delft 28 mei 2024
 
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
Presentatie 8. Joost van der Linde & Daniel Anderton - Eliq 28 mei 2024
 
2024-05-30_meetup_devops_aix-marseille.pdf
2024-05-30_meetup_devops_aix-marseille.pdf2024-05-30_meetup_devops_aix-marseille.pdf
2024-05-30_meetup_devops_aix-marseille.pdf
 
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussion
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussionPro-competitive Industrial Policy – OECD – June 2024 OECD discussion
Pro-competitive Industrial Policy – OECD – June 2024 OECD discussion
 
XP 2024 presentation: A New Look to Leadership
XP 2024 presentation: A New Look to LeadershipXP 2024 presentation: A New Look to Leadership
XP 2024 presentation: A New Look to Leadership
 
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussion
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussionPro-competitive Industrial Policy – LANE – June 2024 OECD discussion
Pro-competitive Industrial Policy – LANE – June 2024 OECD discussion
 
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
Competition and Regulation in Professions and Occupations – OECD – June 2024 ...
 
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussionArtificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
Artificial Intelligence, Data and Competition – OECD – June 2024 OECD discussion
 
ASONAM2023_presection_slide_track-recommendation.pdf
ASONAM2023_presection_slide_track-recommendation.pdfASONAM2023_presection_slide_track-recommendation.pdf
ASONAM2023_presection_slide_track-recommendation.pdf
 
Gregory Harris - Cycle 2 - Civics Presentation
Gregory Harris - Cycle 2 - Civics PresentationGregory Harris - Cycle 2 - Civics Presentation
Gregory Harris - Cycle 2 - Civics Presentation
 
Burning Issue Presentation By Kenmaryon.pdf
Burning Issue Presentation By Kenmaryon.pdfBurning Issue Presentation By Kenmaryon.pdf
Burning Issue Presentation By Kenmaryon.pdf
 
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie WellsCollapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
Collapsing Narratives: Exploring Non-Linearity • a micro report by Rosie Wells
 
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
Suzanne Lagerweij - Influence Without Power - Why Empathy is Your Best Friend...
 
Competition and Regulation in Professions and Occupations – ROBSON – June 202...
Competition and Regulation in Professions and Occupations – ROBSON – June 202...Competition and Regulation in Professions and Occupations – ROBSON – June 202...
Competition and Regulation in Professions and Occupations – ROBSON – June 202...
 
Tom tresser burning issue.pptx My Burning issue
Tom tresser burning issue.pptx My Burning issueTom tresser burning issue.pptx My Burning issue
Tom tresser burning issue.pptx My Burning issue
 
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
原版制作贝德福特大学毕业证(bedfordhire毕业证)硕士文凭原版一模一样
 
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPointMẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
Mẫu PPT kế hoạch làm việc sáng tạo cho nửa cuối năm PowerPoint
 
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
Artificial Intelligence, Data and Competition – SCHREPEL – June 2024 OECD dis...
 
Gregory Harris' Civics Presentation.pptx
Gregory Harris' Civics Presentation.pptxGregory Harris' Civics Presentation.pptx
Gregory Harris' Civics Presentation.pptx
 
Carrer goals.pptx and their importance in real life
Carrer goals.pptx  and their importance in real lifeCarrer goals.pptx  and their importance in real life
Carrer goals.pptx and their importance in real life
 

Content Based Image Retrieval (CBIR)

  • 1. Content Based Image Retrieval (CBIR) Behzad Shomali
  • 2. What is CBIR? Content-based image retrieval, also known as query by image content (QBIC) and content-based visual information retrieval (CBVIR), is the application of computer vision techniques to the image retrieval problem, that is, the problem of searching for digital images in large databases. https://en.wikipedia.org/wiki/Content-based_image_retrieval Query Image Image Feature Extraction Feature Extraction Similarity Measuremen t Retrieved Images
  • 3. Technologies ● Query by example (QBE) ● Semantic retrieval ● Relevance feedback (human interaction) ● Iterative/machine learning ● Other query methods https://en.wikipedia.org/wiki/Content-based_image_retrieval
  • 4. Technologies ● Query by example (QBE) ● Semantic retrieval ● Relevance feedback (human interaction) ● Iterative/machine learning ● Other query methods https://en.wikipedia.org/wiki/Content-based_image_retrieval
  • 5. Application in popular search systems ● Google images ○ Constructing a mathematical model ○ Metadata ● eBay ○ ResNet-50 for category recognition ● SK Planet ○ inception-v3 as vision encoder ○ RNN multi-class classification ● Alibaba ○ GoogLeNet V1 for category prediction and feature learning ● Pinterest ○ Two-step object detection https://en.wikipedia.org/wiki/Reverse_image_search
  • 6. Application in popular search systems ● Google images ○ Constructing a mathematical model ○ Metadata ● eBay ○ ResNet-50 for category recognition ● SK Planet ○ inception-v3 as vision encoder ○ RNN multi-class classification ● Alibaba ○ GoogLeNet V1 for category prediction and feature learning ● Pinterest ○ Two-step object detection https://en.wikipedia.org/wiki/Reverse_image_search
  • 7. Image Representation and Features ● Extract local and deep features ● Studied AlexNet and VGG ○ Extract feature representations from fc6 and fc8 layers ○ Binarized ○ Hamming distance ● Extract salient color signatures ○ Detect salient regions ○ K-means clustering ○ Store cluster centroids and weights as image signature [DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
  • 8. Two-step Object Detection and Localization 1. Category classification 2. Object detection [DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
  • 9. Car Flower Person c1 ... cn f1 ... fm p1 ... pk Input Category classification Object detection
  • 10. Car Flower Person c1 ... cn f1 ... fm p1 ... pk Input Category classification Object detection
  • 11. Car Flower Person c1 ... cn f1 ... fm p1 ... pk Input Category classification Object detection Reduce computational cost
  • 12. Static Evaluation of Search Relevance ● Used dataset contains: 1.6 M unique images ○ Be assumed to be relevant, if two images share a label ● Computed precision@k based on several features ○ The fc6 layer activations from the generic AlexNet (pre-trained for ILSVRC) ○ The fc6 activations of an AlexNet model fine-tuned to recognize over 3,000 Pinterest products categories ○ The loss3/classifier activations of a generic GoogLeNet ○ The fc6 activations of a generic VGG 16-layer model [DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
  • 13. Precision vs. Recall [Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601]
  • 14. Precision vs. Recall [Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601] Either value alone contains insufficient information ● We can always make recall 1, simply by retrieving all images ● Similarly, precision can be kept high by retrieving only a few images ● P (10) ; P (30) ; P (NR) - the precision after the first 10 ; 30 ; NR documents are retrieved ● Mean Average Precision - mean (non-interpolated) average precision . ● recall at .5 precision - recall at the rank where precision drops below .5. ● R (1000) - recall after 1000 documents are retrieved. ● Rank first relevant - The rank of the highest-ranked relevant document.
  • 15. Precision vs. Recall [Müller, Henning, et al. "Performance evaluation in content-based image retrieval: overview and proposals." Pattern recognition letters 22.5 (2001): 593-601] Either value alone contains insufficient information ● We can always make recall 1, simply by retrieving all images ● Similarly, precision can be kept high by retrieving only a few images Precision and recall should either be used together ● P (10) ; P (30) ; P (NR) - the precision after the first 10 ; 30 ; NR documents are retrieved ● Mean Average Precision - mean (non-interpolated) average precision . ● recall at .5 precision - recall at the rank where precision drops below .5. ● R (1000) - recall after 1000 documents are retrieved. ● Rank first relevant - The rank of the highest-ranked relevant document.
  • 16. Relevance of visual search Table 1 shows p@5 and p@10 performance of these models, along with the average CPU-based latency of our visual search service, which includes feature extraction for the query image as well as retrieval. [DavJing, Yushi, et al. "Visual search at pinterest." Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 2015]
  • 17. Siamese networks [Das, Arpita, et al. "Together we stand: Siamese networks for similar question retrieval." Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2016]
  • 18. Siamese networks ● Let, F(X) be the family of functions with set of parameters W. F(X) is assumed to be differentiable with respect to W. Siamese network seeks a value of the parameter W such that the symmetric similarity metric is small if X1 and X2 belong to the same category, and large if they belong to different categories. [Das, Arpita, et al. "Together we stand: Siamese networks for similar question retrieval." Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2016]
  • 19. Different loss functions for training a Siamese network Two commonly used ones are ● Triplet loss ● Contrastive loss The main idea of these loss functions is to pull the samples of every class toward one another and push the samples of different classes away from each other [Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural Networks (IJCNN). IEEE, 2020]
  • 20. Different loss functions - Triplet loss The triplet loss uses anchor, neighbor, and distant. Let f(x) be the output (i.e., embedding) of the network for the input x. The triplet loss tries to reduce the distance of anchor and neighbor embeddings and desires to increase the distance of anchor and distant embeddings. As long as the distances of anchor-distant pairs get larger than the distances of anchor-neighbor pairs by a margin α ≥ 0, the desired embedding is obtained [Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural Networks (IJCNN). IEEE, 2020]
  • 21. Different loss functions - Contrastive loss The contrastive loss uses pairs of samples which can be anchor and neighbor or anchor and distant. If the samples are anchor and neighbor, they are pulled towards each other; otherwise, their distance is increased. In other words, the contrastive loss performs like the triplet loss but one by one rather than simultaneously. The desired embedding is obtained when the anchor-distant distances get larger than the anchor-neighbor distances by a margin of α [Ghojogh, Benyamin, et al. "Fisher discriminant triplet and contrastive losses for training siamese networks." 2020 International Joint Conference on Neural Networks (IJCNN). IEEE, 2020]