SlideShare a Scribd company logo
1 of 17
Mumford-Shah Loss Functional for Image
Segmentation with Deep Learning
IEEE Transactions on Image Processing
Boah Kim, Jong Chul Ye
 Fully supervised segmentation : 100% pixel-wise labels
 Semi-supervised segmentation: <100% of pixel-wise labels
 Weakly-supervised segmentation: No pixel-wise labels, rather Image-level / bounding box labels
 Unsupervised segmentation: No labels
Introduction
Semi-, Un- supervised Image Segmentation
No label
Image Segmentation
 Fully pixel-wise labels : time-consuming, difficult to obtain in certain domains (ex. medical images)
 How to use unlabeled images without any (image-level, bounding box) labels?
⇒ Revisiting the classical image segmentation method, “Mumford-Shah functional”
Introduction
Semi-, Un- supervised Image Segmentation
No label
Image Segmentation
Related works
Semi-, Un- supervised Image Segmentation
Mumford-Shah Functional
- 1st term : Distance between the model
and the input image
- 2nd term : Smoothness of the model
within the sub-regions
 Optimality criterion for segmenting an image into sub-regions
 Chan-Vese, multiphase level-set framework
- Make the characteristic function be differentiable with Heaviside function
Main Theory
Semi-, Un- supervised Image Segmentation
Mumford-Shah Loss Functional
 Similarity between the characteristic function and softmax layer in CNN
 Proposed “Mumford-Shah loss functional”
Main Theory
Semi-, Un- supervised Image Segmentation
Application of Mumford-Shah Loss
 In the presence of semantic labels,
Main Theory
Semi-, Un- supervised Image Segmentation
Application of Mumford-Shah Loss
 In the absence of semantic labels,
Main Theory
Semi-, Un- supervised Image Segmentation
Application of Mumford-Shah Loss
 In the absence of semantic labels,
Main Theory
Semi-, Un- supervised Image Segmentation
Application of Mumford-Shah Loss
 In the absence of semantic labels,
- For image with intensity inhomogeneities, impose a term of bias field estimation
Experiments
Semi-, Un- supervised Image Segmentation
 Semi-supervised object segmentation in Natural Images
- PASCAL VOC 2012 dataset
 Semi-supervised tumor segmentation in Medical Images
- LiTS 2017 dataset
- BRATS 2015 dataset
 Unsupervised segmentation in Natural images
- BSDS500 dataset
- PASCAL VOC 2012 dataset
Experiments
Semi-, Un- supervised Image Segmentation
Semi-supervised Object Segmentation in Natural Images
 Qualitative evaluation on PASCAL VOC 2012 dataset
Experiments
Semi-, Un- supervised Image Segmentation
Semi-supervised Object Segmentation in Natural Images
 Quantitative evaluation on PASCAL VOC 2012 dataset (Training with ¼ labeled data)
Experiments
Semi-, Un- supervised Image Segmentation
Semi-supervised Tumor Segmentation in Medical Images
 Qualitative evaluation on LiTS and BRATS datasets
1/10 labeled data on LiTS dataset 1/4 labeled data on BRATS dataset
Experiments
Semi-, Un- supervised Image Segmentation
Semi-supervised Tumor Segmentation in Medical Images
 Quantitative evaluation on LiTS and BRATS datasets
LiTS dataset
BRATS dataset
(1/4 labeled data)
Experiments
Semi-, Un- supervised Image Segmentation
Unsupervised Segmentation in Natural Images
 Evaluation on BSDS500 dataset
Experiments
Semi-, Un- supervised Image Segmentation
Unsupervised Segmentation in Natural Images
 Evaluation on PASCAL VOC 2012 dataset
Thank you.

More Related Content

What's hot

UNetEliyaLaialy (2).pptx
UNetEliyaLaialy (2).pptxUNetEliyaLaialy (2).pptx
UNetEliyaLaialy (2).pptx
NoorUlHaq47
 
Image segmentation
Image segmentationImage segmentation
Image segmentation
Deepak Kumar
 
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation..."Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
Edge AI and Vision Alliance
 
Enhancement in frequency domain
Enhancement in frequency domainEnhancement in frequency domain
Enhancement in frequency domain
Ashish Kumar
 

What's hot (20)

Super Resolution
Super ResolutionSuper Resolution
Super Resolution
 
UNetEliyaLaialy (2).pptx
UNetEliyaLaialy (2).pptxUNetEliyaLaialy (2).pptx
UNetEliyaLaialy (2).pptx
 
Image segmentation
Image segmentationImage segmentation
Image segmentation
 
Computer Vision Structure from motion
Computer Vision Structure from motionComputer Vision Structure from motion
Computer Vision Structure from motion
 
The single image dehazing based on efficient transmission estimation
The single image dehazing based on efficient transmission estimationThe single image dehazing based on efficient transmission estimation
The single image dehazing based on efficient transmission estimation
 
Grey-level Co-occurence features for salt texture classification
Grey-level Co-occurence features for salt texture classificationGrey-level Co-occurence features for salt texture classification
Grey-level Co-occurence features for salt texture classification
 
Image segmentation
Image segmentationImage segmentation
Image segmentation
 
ImageProcessing10-Segmentation(Thresholding) (1).ppt
ImageProcessing10-Segmentation(Thresholding) (1).pptImageProcessing10-Segmentation(Thresholding) (1).ppt
ImageProcessing10-Segmentation(Thresholding) (1).ppt
 
Segmentation
SegmentationSegmentation
Segmentation
 
Feature Matching using SIFT algorithm
Feature Matching using SIFT algorithmFeature Matching using SIFT algorithm
Feature Matching using SIFT algorithm
 
Image segmentation
Image segmentationImage segmentation
Image segmentation
 
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation..."Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
"Semantic Segmentation for Scene Understanding: Algorithms and Implementation...
 
Single Image Super Resolution Overview
Single Image Super Resolution OverviewSingle Image Super Resolution Overview
Single Image Super Resolution Overview
 
Image compression in digital image processing
Image compression in digital image processingImage compression in digital image processing
Image compression in digital image processing
 
Image segmentation
Image segmentationImage segmentation
Image segmentation
 
Enhancement in frequency domain
Enhancement in frequency domainEnhancement in frequency domain
Enhancement in frequency domain
 
Edge linking in image processing
Edge linking in image processingEdge linking in image processing
Edge linking in image processing
 
JPEG Image Compression
JPEG Image CompressionJPEG Image Compression
JPEG Image Compression
 
Image Enhancement in Spatial Domain
Image Enhancement in Spatial DomainImage Enhancement in Spatial Domain
Image Enhancement in Spatial Domain
 
Lect 02 second portion
Lect 02  second portionLect 02  second portion
Lect 02 second portion
 

Similar to Mumford-Shah Loss Functional for Image Segmentation With Deep Learning

research paper Ijetae 0812 23
research paper Ijetae 0812 23research paper Ijetae 0812 23
research paper Ijetae 0812 23
Punit Karnani
 
Automated quality assurance in whole slide pathology images blurred region de...
Automated quality assurance in whole slide pathology images blurred region de...Automated quality assurance in whole slide pathology images blurred region de...
Automated quality assurance in whole slide pathology images blurred region de...
BIT002
 
Medical Image Processing � Detection of Cancer Brain
Medical Image Processing � Detection of Cancer BrainMedical Image Processing � Detection of Cancer Brain
Medical Image Processing � Detection of Cancer Brain
ijcnes
 
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
Editor IJMTER
 

Similar to Mumford-Shah Loss Functional for Image Segmentation With Deep Learning (20)

LEARNING BASES OF ACTICITY
LEARNING BASES OF ACTICITYLEARNING BASES OF ACTICITY
LEARNING BASES OF ACTICITY
 
wepik-optimizing-image-segmentation-through-multiple-threshold-analysis-20240...
wepik-optimizing-image-segmentation-through-multiple-threshold-analysis-20240...wepik-optimizing-image-segmentation-through-multiple-threshold-analysis-20240...
wepik-optimizing-image-segmentation-through-multiple-threshold-analysis-20240...
 
Utilizing image scales towards totally training free blind image quality asse...
Utilizing image scales towards totally training free blind image quality asse...Utilizing image scales towards totally training free blind image quality asse...
Utilizing image scales towards totally training free blind image quality asse...
 
Remote Sensing Image Scene Classification
Remote Sensing Image Scene ClassificationRemote Sensing Image Scene Classification
Remote Sensing Image Scene Classification
 
A feature enriched completely blind image
A feature enriched completely blind imageA feature enriched completely blind image
A feature enriched completely blind image
 
Whitepaper: Does Image Quality Matter?
Whitepaper: Does Image Quality Matter?Whitepaper: Does Image Quality Matter?
Whitepaper: Does Image Quality Matter?
 
research paper Ijetae 0812 23
research paper Ijetae 0812 23research paper Ijetae 0812 23
research paper Ijetae 0812 23
 
Image annotation - Segmentation & Annotation
Image annotation - Segmentation & AnnotationImage annotation - Segmentation & Annotation
Image annotation - Segmentation & Annotation
 
[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...
[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...
[MICCAI 2022] Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Imag...
 
IRJET-Semi-Supervised Collaborative Image Retrieval using Relevance Feedback
IRJET-Semi-Supervised Collaborative Image Retrieval using Relevance FeedbackIRJET-Semi-Supervised Collaborative Image Retrieval using Relevance Feedback
IRJET-Semi-Supervised Collaborative Image Retrieval using Relevance Feedback
 
Ko3419161921
Ko3419161921Ko3419161921
Ko3419161921
 
A Survey on Different Relevance Feedback Techniques in Content Based Image Re...
A Survey on Different Relevance Feedback Techniques in Content Based Image Re...A Survey on Different Relevance Feedback Techniques in Content Based Image Re...
A Survey on Different Relevance Feedback Techniques in Content Based Image Re...
 
MEDICAL IMAGE TEXTURE SEGMENTATION USINGRANGE FILTER
MEDICAL IMAGE TEXTURE SEGMENTATION USINGRANGE FILTERMEDICAL IMAGE TEXTURE SEGMENTATION USINGRANGE FILTER
MEDICAL IMAGE TEXTURE SEGMENTATION USINGRANGE FILTER
 
Automated quality assurance in whole slide pathology images blurred region de...
Automated quality assurance in whole slide pathology images blurred region de...Automated quality assurance in whole slide pathology images blurred region de...
Automated quality assurance in whole slide pathology images blurred region de...
 
Literature Review on Content Based Image Retrieval
Literature Review on Content Based Image RetrievalLiterature Review on Content Based Image Retrieval
Literature Review on Content Based Image Retrieval
 
Person Recognition
Person RecognitionPerson Recognition
Person Recognition
 
International Journal of Computational Engineering Research(IJCER)
 International Journal of Computational Engineering Research(IJCER)  International Journal of Computational Engineering Research(IJCER)
International Journal of Computational Engineering Research(IJCER)
 
Dip fundamentals 2
Dip fundamentals 2Dip fundamentals 2
Dip fundamentals 2
 
Medical Image Processing � Detection of Cancer Brain
Medical Image Processing � Detection of Cancer BrainMedical Image Processing � Detection of Cancer Brain
Medical Image Processing � Detection of Cancer Brain
 
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
FACE EXPRESSION IDENTIFICATION USING IMAGE FEATURE CLUSTRING AND QUERY SCHEME...
 

More from BoahKim2

More from BoahKim2 (8)

ICLR2023_DARL.pdf
ICLR2023_DARL.pdfICLR2023_DARL.pdf
ICLR2023_DARL.pdf
 
TIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdfTIP_TAViT_presentation.pdf
TIP_TAViT_presentation.pdf
 
DiffuseMorph: Unsupervised Deformable Image Registration Using Diffusion Model
DiffuseMorph: Unsupervised Deformable Image Registration Using Diffusion ModelDiffuseMorph: Unsupervised Deformable Image Registration Using Diffusion Model
DiffuseMorph: Unsupervised Deformable Image Registration Using Diffusion Model
 
Diffusion Deformable Model for 4D Temporal Medical Image Generation
Diffusion Deformable Model for 4D Temporal Medical Image GenerationDiffusion Deformable Model for 4D Temporal Medical Image Generation
Diffusion Deformable Model for 4D Temporal Medical Image Generation
 
Unsupervised Deformable Image Registration Using Cycle-Consistent CNN
Unsupervised Deformable Image Registration Using Cycle-Consistent CNNUnsupervised Deformable Image Registration Using Cycle-Consistent CNN
Unsupervised Deformable Image Registration Using Cycle-Consistent CNN
 
CycleMorph: Cycle consistent unsupervised deformable image registration
CycleMorph: Cycle consistent unsupervised deformable image registrationCycleMorph: Cycle consistent unsupervised deformable image registration
CycleMorph: Cycle consistent unsupervised deformable image registration
 
Medical Image Synthesis with Improved Cycle-GAN: CT from CECT
Medical Image Synthesis with Improved Cycle-GAN: CT from CECT Medical Image Synthesis with Improved Cycle-GAN: CT from CECT
Medical Image Synthesis with Improved Cycle-GAN: CT from CECT
 
Multi-energy Bone Subtraction in Chest Radiography by Eigenvalue Decomposition
Multi-energy Bone Subtraction in Chest Radiography by Eigenvalue DecompositionMulti-energy Bone Subtraction in Chest Radiography by Eigenvalue Decomposition
Multi-energy Bone Subtraction in Chest Radiography by Eigenvalue Decomposition
 

Recently uploaded

Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak HamilCara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Kandungan 087776558899
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
Kamal Acharya
 
Digital Communication Essentials: DPCM, DM, and ADM .pptx
Digital Communication Essentials: DPCM, DM, and ADM .pptxDigital Communication Essentials: DPCM, DM, and ADM .pptx
Digital Communication Essentials: DPCM, DM, and ADM .pptx
pritamlangde
 
Introduction to Robotics in Mechanical Engineering.pptx
Introduction to Robotics in Mechanical Engineering.pptxIntroduction to Robotics in Mechanical Engineering.pptx
Introduction to Robotics in Mechanical Engineering.pptx
hublikarsn
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
Epec Engineered Technologies
 

Recently uploaded (20)

Design For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the startDesign For Accessibility: Getting it right from the start
Design For Accessibility: Getting it right from the start
 
Memory Interfacing of 8086 with DMA 8257
Memory Interfacing of 8086 with DMA 8257Memory Interfacing of 8086 with DMA 8257
Memory Interfacing of 8086 with DMA 8257
 
Introduction to Geographic Information Systems
Introduction to Geographic Information SystemsIntroduction to Geographic Information Systems
Introduction to Geographic Information Systems
 
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak HamilCara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
Cara Menggugurkan Sperma Yang Masuk Rahim Biyar Tidak Hamil
 
Hospital management system project report.pdf
Hospital management system project report.pdfHospital management system project report.pdf
Hospital management system project report.pdf
 
Online food ordering system project report.pdf
Online food ordering system project report.pdfOnline food ordering system project report.pdf
Online food ordering system project report.pdf
 
Hostel management system project report..pdf
Hostel management system project report..pdfHostel management system project report..pdf
Hostel management system project report..pdf
 
Employee leave management system project.
Employee leave management system project.Employee leave management system project.
Employee leave management system project.
 
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
COST-EFFETIVE  and Energy Efficient BUILDINGS ptxCOST-EFFETIVE  and Energy Efficient BUILDINGS ptx
COST-EFFETIVE and Energy Efficient BUILDINGS ptx
 
Digital Communication Essentials: DPCM, DM, and ADM .pptx
Digital Communication Essentials: DPCM, DM, and ADM .pptxDigital Communication Essentials: DPCM, DM, and ADM .pptx
Digital Communication Essentials: DPCM, DM, and ADM .pptx
 
Introduction to Robotics in Mechanical Engineering.pptx
Introduction to Robotics in Mechanical Engineering.pptxIntroduction to Robotics in Mechanical Engineering.pptx
Introduction to Robotics in Mechanical Engineering.pptx
 
Introduction to Artificial Intelligence ( AI)
Introduction to Artificial Intelligence ( AI)Introduction to Artificial Intelligence ( AI)
Introduction to Artificial Intelligence ( AI)
 
Post office management system project ..pdf
Post office management system project ..pdfPost office management system project ..pdf
Post office management system project ..pdf
 
Standard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power PlayStandard vs Custom Battery Packs - Decoding the Power Play
Standard vs Custom Battery Packs - Decoding the Power Play
 
Introduction to Serverless with AWS Lambda
Introduction to Serverless with AWS LambdaIntroduction to Serverless with AWS Lambda
Introduction to Serverless with AWS Lambda
 
Theory of Time 2024 (Universal Theory for Everything)
Theory of Time 2024 (Universal Theory for Everything)Theory of Time 2024 (Universal Theory for Everything)
Theory of Time 2024 (Universal Theory for Everything)
 
Electromagnetic relays used for power system .pptx
Electromagnetic relays used for power system .pptxElectromagnetic relays used for power system .pptx
Electromagnetic relays used for power system .pptx
 
School management system project Report.pdf
School management system project Report.pdfSchool management system project Report.pdf
School management system project Report.pdf
 
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptxS1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
S1S2 B.Arch MGU - HOA1&2 Module 3 -Temple Architecture of Kerala.pptx
 
NO1 Top No1 Amil Baba In Azad Kashmir, Kashmir Black Magic Specialist Expert ...
NO1 Top No1 Amil Baba In Azad Kashmir, Kashmir Black Magic Specialist Expert ...NO1 Top No1 Amil Baba In Azad Kashmir, Kashmir Black Magic Specialist Expert ...
NO1 Top No1 Amil Baba In Azad Kashmir, Kashmir Black Magic Specialist Expert ...
 

Mumford-Shah Loss Functional for Image Segmentation With Deep Learning

  • 1. Mumford-Shah Loss Functional for Image Segmentation with Deep Learning IEEE Transactions on Image Processing Boah Kim, Jong Chul Ye
  • 2.  Fully supervised segmentation : 100% pixel-wise labels  Semi-supervised segmentation: <100% of pixel-wise labels  Weakly-supervised segmentation: No pixel-wise labels, rather Image-level / bounding box labels  Unsupervised segmentation: No labels Introduction Semi-, Un- supervised Image Segmentation No label Image Segmentation
  • 3.  Fully pixel-wise labels : time-consuming, difficult to obtain in certain domains (ex. medical images)  How to use unlabeled images without any (image-level, bounding box) labels? ⇒ Revisiting the classical image segmentation method, “Mumford-Shah functional” Introduction Semi-, Un- supervised Image Segmentation No label Image Segmentation
  • 4. Related works Semi-, Un- supervised Image Segmentation Mumford-Shah Functional - 1st term : Distance between the model and the input image - 2nd term : Smoothness of the model within the sub-regions  Optimality criterion for segmenting an image into sub-regions  Chan-Vese, multiphase level-set framework - Make the characteristic function be differentiable with Heaviside function
  • 5. Main Theory Semi-, Un- supervised Image Segmentation Mumford-Shah Loss Functional  Similarity between the characteristic function and softmax layer in CNN  Proposed “Mumford-Shah loss functional”
  • 6. Main Theory Semi-, Un- supervised Image Segmentation Application of Mumford-Shah Loss  In the presence of semantic labels,
  • 7. Main Theory Semi-, Un- supervised Image Segmentation Application of Mumford-Shah Loss  In the absence of semantic labels,
  • 8. Main Theory Semi-, Un- supervised Image Segmentation Application of Mumford-Shah Loss  In the absence of semantic labels,
  • 9. Main Theory Semi-, Un- supervised Image Segmentation Application of Mumford-Shah Loss  In the absence of semantic labels, - For image with intensity inhomogeneities, impose a term of bias field estimation
  • 10. Experiments Semi-, Un- supervised Image Segmentation  Semi-supervised object segmentation in Natural Images - PASCAL VOC 2012 dataset  Semi-supervised tumor segmentation in Medical Images - LiTS 2017 dataset - BRATS 2015 dataset  Unsupervised segmentation in Natural images - BSDS500 dataset - PASCAL VOC 2012 dataset
  • 11. Experiments Semi-, Un- supervised Image Segmentation Semi-supervised Object Segmentation in Natural Images  Qualitative evaluation on PASCAL VOC 2012 dataset
  • 12. Experiments Semi-, Un- supervised Image Segmentation Semi-supervised Object Segmentation in Natural Images  Quantitative evaluation on PASCAL VOC 2012 dataset (Training with ¼ labeled data)
  • 13. Experiments Semi-, Un- supervised Image Segmentation Semi-supervised Tumor Segmentation in Medical Images  Qualitative evaluation on LiTS and BRATS datasets 1/10 labeled data on LiTS dataset 1/4 labeled data on BRATS dataset
  • 14. Experiments Semi-, Un- supervised Image Segmentation Semi-supervised Tumor Segmentation in Medical Images  Quantitative evaluation on LiTS and BRATS datasets LiTS dataset BRATS dataset (1/4 labeled data)
  • 15. Experiments Semi-, Un- supervised Image Segmentation Unsupervised Segmentation in Natural Images  Evaluation on BSDS500 dataset
  • 16. Experiments Semi-, Un- supervised Image Segmentation Unsupervised Segmentation in Natural Images  Evaluation on PASCAL VOC 2012 dataset