Diabetic Foot Ulcers Grand Challenge: Second Challenge, DFUC 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, September 27, 2021, ...

Preface
Organization
Contents
Development of Diabetic Foot Ulcer Datasets: An Overview
	1 Introduction
	2 A Survey of Data Capturing Methods
	3 A Review of DFU Image Datasets
		3.1 Binary Classification
		3.2 DFU Detection (DFUC2020)
		3.3 Multi-class DFU Classification (DFUC2021)
		3.4 Other Datasets
	4 DFU Challenges
	5 Future Directions
	6 Conclusion
	References
DFUC 2021 Challenge Papers
Convolutional Nets Versus Vision Transformers for Diabetic Foot Ulcer Classification
	1 Introduction and Problem Statement
	2 Methodology
		2.1 Convolutional Neural Networks and Vision Transformers
		2.2 Sharpness-Aware Optimization and Training Details
		2.3 Training Data
	3 Experimental Results and Analysis
		3.1 Impact of SAM Optimization
		3.2 CNNs vs Transformers
		3.3 Discussion of the Results
	4 Conclusion
	References
Boosting EfficientNets Ensemble Performance via Pseudo-Labels and Synthetic Images by pix2pixHD for Infection and Ischaemia Classification in Diabetic Foot Ulcers
	1 Introduction
	2 Data and Methods
		2.1 Diabetic Foot Ulcer Challenge 2021 Dataset
		2.2 Classification via EfficientNets
		2.3 Image Synthesis via pix2pixHD
		2.4 Experimental Environment
	3 Approach
		3.1 Baseline Models and Prediction Ensemble
		3.2 Pseudo-Labeling and Synthetic Image Generation
		3.3 Extended Models and Prediction Ensemble
	4 Results
		4.1 Baseline Model and Ensemble Performance
		4.2 Synthetic Images for Training Dataset Extension
		4.3 Extended Model and Ensemble Performances
		4.4 Local Interpretable Model-Agnostic Explanations (LIME)
	5 Discussion
		5.1 Models and Ensembles
		5.2 Pseudo-Labeling and Synthetic Image Generation
		5.3 Limitations
	6 Conclusion
	References
Bias Adjustable Activation Network for Imbalanced Data—Diabetic Foot Ulcer Challenge 2021
	1 Introduction
	2 Methodology
		2.1 Data Augmentation
		2.2 Scheduler Adjustments
		2.3 Weighted Cross Entropy
		2.4 Bias-Adjustable Softmax
		2.5 Inference Pipeline
	3 Results and Discussion
	4 Conclusion
	References
Efficient Multi-model Vision Transformer Based on Feature Fusion for Classification of DFUC2021 Challenge
	1 Introduction
	2 Material and Methods
		2.1 Dataset
		2.2 Related Work
		2.3 Proposed Method
	3 Performance Metrics
	4 Results and Discussion
	5 Conclusion
	References
Classification of Infection and Ischemia in Diabetic Foot Ulcers Using VGG Architectures
	1 Introduction
	2 Materials and Methods
		2.1 DFU Dataset
		2.2 DL Framework
		2.3 Architecture
		2.4 Training
	3 Results
		3.1 Training/Validation Dataset Performance
		3.2 Testing Dataset Performance
	4 Discussion
	5 Conclusions
	A  Illustrations of Various VGG Variants
		A.1  VGG16
		A.2  VGG19
	References
Diabetic Foot Ulcer Grand Challenge 2021: Evaluation and Summary
	1 Introduction
	2 Methodology
		2.1 Datasets and Ground Truth
		2.2 Performance Metrics
		2.3 Analysis of the Proposed Methods
	3 Results and Discussion
		3.1 Analysis on the Top-3 Results
		3.2 Ensemble of the Top 10 Performing Models
		3.3 Visual Comparison of the Top-10 Methods
	4 Conclusion
	References
Post Challenge Paper
Deep Subspace Analysing for Semi-supervised Multi-label Classification of Diabetic Foot Ulcer
	1 Introduction
	2 DFUC2021
	3 Methodology
		3.1 Stage 1: Transfer Learning
		3.2 Stage 2: Deep Subspace-Based Descriptors
		3.3 Stage 3: Final Classification
	4 Results and Discussion
	5 Conclusion
	References
Author Index