Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Universal Semi-Supervised Learning for Medical Image Classification

Lie Ju, Yicheng Wu, Wei Feng, Zhen Yu, Lin Wang, Zhuoting Zhu, Zongyuan Ge

TL;DR

This work tackles open-set and cross-domain challenges in medical image SSL by introducing a universal SSL framework that jointly detects unseen classes (UKC) and unseen domains (UKD) and then leverages them through domain adaptation and Pi-model SSL. A dual-path outlier estimation (DOE) uses class prototypes and prediction agreement to score UKC samples, while a VAE-based domain separation with a two-component GMM assigns UKD relevance for targeted adaptation. The approach integrates adversarial training with SSL, guided by reweighting from both DOE and CDS, and demonstrates superior performance on dermatology and ophthalmology tasks under various open-set conditions. The results suggest practical benefit for real-world medical imaging where unlabeled data come from diverse, unfamiliar sources.

Abstract

Semi-supervised learning (SSL) has attracted much attention since it reduces the expensive costs of collecting adequate well-labeled training data, especially for deep learning methods. However, traditional SSL is built upon an assumption that labeled and unlabeled data should be from the same distribution \textit{e.g.,} classes and domains. However, in practical scenarios, unlabeled data would be from unseen classes or unseen domains, and it is still challenging to exploit them by existing SSL methods. Therefore, in this paper, we proposed a unified framework to leverage these unseen unlabeled data for open-scenario semi-supervised medical image classification. We first design a novel scoring mechanism, called dual-path outliers estimation, to identify samples from unseen classes. Meanwhile, to extract unseen-domain samples, we then apply an effective variational autoencoder (VAE) pre-training. After that, we conduct domain adaptation to fully exploit the value of the detected unseen-domain samples to boost semi-supervised training. We evaluated our proposed framework on dermatology and ophthalmology tasks. Extensive experiments demonstrate our model can achieve superior classification performance in various medical SSL scenarios. The code implementations are accessible at: https://github.com/PyJulie/USSL4MIC.

Universal Semi-Supervised Learning for Medical Image Classification

TL;DR

This work tackles open-set and cross-domain challenges in medical image SSL by introducing a universal SSL framework that jointly detects unseen classes (UKC) and unseen domains (UKD) and then leverages them through domain adaptation and Pi-model SSL. A dual-path outlier estimation (DOE) uses class prototypes and prediction agreement to score UKC samples, while a VAE-based domain separation with a two-component GMM assigns UKD relevance for targeted adaptation. The approach integrates adversarial training with SSL, guided by reweighting from both DOE and CDS, and demonstrates superior performance on dermatology and ophthalmology tasks under various open-set conditions. The results suggest practical benefit for real-world medical imaging where unlabeled data come from diverse, unfamiliar sources.

Abstract

Semi-supervised learning (SSL) has attracted much attention since it reduces the expensive costs of collecting adequate well-labeled training data, especially for deep learning methods. However, traditional SSL is built upon an assumption that labeled and unlabeled data should be from the same distribution \textit{e.g.,} classes and domains. However, in practical scenarios, unlabeled data would be from unseen classes or unseen domains, and it is still challenging to exploit them by existing SSL methods. Therefore, in this paper, we proposed a unified framework to leverage these unseen unlabeled data for open-scenario semi-supervised medical image classification. We first design a novel scoring mechanism, called dual-path outliers estimation, to identify samples from unseen classes. Meanwhile, to extract unseen-domain samples, we then apply an effective variational autoencoder (VAE) pre-training. After that, we conduct domain adaptation to fully exploit the value of the detected unseen-domain samples to boost semi-supervised training. We evaluated our proposed framework on dermatology and ophthalmology tasks. Extensive experiments demonstrate our model can achieve superior classification performance in various medical SSL scenarios. The code implementations are accessible at: https://github.com/PyJulie/USSL4MIC.
Paper Structure (13 sections, 8 equations, 3 figures, 4 tables)

This paper contains 13 sections, 8 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overview
  4. Dual-path Outlier Estimation
  5. Class-agnostic Domain Separation
  6. Optimization with Adversarial Training & SSL
  7. Experiments
  8. Datasets & Implementation Details
  9. Comparison Study
  10. Ablation Study
  11. Conclusion
  12. Acknowledgments.
  13. Disclosure of Interests.

Figures (3)

  • Figure 1: Problem illustration. (a) Close-set SSL. The samples in the labeled and unlabeled data share the same classes and are collected under the same environment, i.e., dermatoscopes. (b) Open-set SSL. There are unknown classes (UKC) in the unlabeled data, e.g., BCC and BKL. (c) Universal SSL. In addition to the unknown classes, the samples in the unlabeled data may come from other unknown domains (UKD), e.g., samples from other datasets with different imaging and condition settings.
  • Figure 2: The overview of our proposed framework.
  • Figure 3: The visualized examples from unlabeled data with normalized scores.