Robustly Optimized Deep Feature Decoupling Network for Fatty Liver Diseases Detection

TL;DR

This paper introduces a robust deep learning framework for fatty liver disease detection that decouples features and uses adaptive adversarial training to improve classification accuracy and robustness, especially for underperforming classes.

Contribution

The proposed method combines feature decoupling with adaptive adversarial training, enhancing class balance and robustness in fatty liver classification without requiring massive data.

Findings

Accuracy improved by 4.16% to 82.95%

Effective in eliminating recognition weaknesses

Generalized framework applicable to other classifiers

Abstract

Current medical image classification efforts mainly aim for higher average performance, often neglecting the balance between different classes. This can lead to significant differences in recognition accuracy between classes and obvious recognition weaknesses. Without the support of massive data, deep learning faces challenges in fine-grained classification of fatty liver. In this paper, we propose an innovative deep learning framework that combines feature decoupling and adaptive adversarial training. Firstly, we employ two iteratively compressed decouplers to supervised decouple common features and specific features related to fatty liver in abdominal ultrasound images. Subsequently, the decoupled features are concatenated with the original image after transforming the color space and are fed into the classifier. During adversarial training, we adaptively adjust the perturbation and balance the adversarial strength by the accuracy of each class. The model will eliminate recognition weaknesses by correctly classifying adversarial samples, thus improving recognition robustness. Finally, the accuracy of our method improved by 4.16%, achieving 82.95%. As demonstrated by extensive experiments, our method is a generalized learning framework that can be directly used to eliminate the recognition weaknesses of any classifier while improving its average performance. Code is available at https://github.com/HP-ML/MICCAI2024.