Improving DFU Image Classification by an Adaptive Augmentation Pool and Voting with Expertise
Authors:
- 21 April 2023
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 196-202
Abstract
Medical images are inseparable from data augmentation due to the lack of labeled data. However, several existing effective augments for natural images may not be suitable for medical images. This paper proposes an adaptive framework to construct a suitable augmentation pool for Diabetic Foot Ulcers medical images. Additionally, we use ensemble learning to enhance the model’s output. Instead of commonly used plurality voting, we propose a strategy named “voting with expertise” which prioritizes prediction with adequately reliable value. Experimental results show the effectiveness of the proposed methods and we won second place by combining the above two improvements in the ongoing open challenge-DFUC2021 Challenge.Keywords
Funding Information
- Research and Development
- National Natural Science Foundation of China
This publication has 16 references indexed in Scilit:
- A review of medical image data augmentation techniques for deep learning applicationsJournal of Medical Imaging and Radiation Oncology, 2021
- Image Augmentation for Deep Learning Based Lesion Classification from Skin Images2020
- Group NormalizationInternational Journal of Computer Vision, 2019
- A survey on Image Data Augmentation for Deep LearningJournal of Big Data, 2019
- DFU_QUTNet: diabetic foot ulcer classification using novel deep convolutional neural networkMultimedia Tools and Applications, 2019
- DFUNet: Convolutional Neural Networks for Diabetic Foot Ulcer ClassificationIEEE Transactions on Emerging Topics in Computational Intelligence, 2018
- Data augmentation for improving deep learning in image classification problem2018
- Diabetic Foot Ulcers and Their RecurrenceNew England Journal of Medicine, 2017
- Identity Mappings in Deep Residual Networks2016
- Ensemble Methods in Machine Learning2000