TagGAN: A Generative Model for Data Tagging

2025-02-25

AI Summary

Medical image analysis urgently requires interpretable, pixel-level lesion localization, yet existing methods rely on costly pixel-level annotations. This paper proposes a weakly supervised, interpretable domain adaptation framework that generates fine-grained disease heatmaps and binary lesion masks using only image-level labels. Our approach addresses the problem of annotation scarcity by enabling high-fidelity lesion localization without any pixel-level supervision. Methodologically, we introduce an interpretability-driven GAN architecture integrating feature disentanglement, residual mapping, and lesion-aware loss functions, coupled with semantically consistent domain translation from abnormal to normal representations. Experimentally, our method achieves state-of-the-art performance on CheXpert, TBX11K, and COVID-19 datasets, significantly improving lesion pixel identification accuracy while drastically reducing clinical annotation burden. Key contributions include: (1) the first pixel-accurate lesion localization under pure image-level supervision; (2) a novel GAN design explicitly optimized for interpretability; and (3) effective semantic domain alignment enabling robust generalization.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Precise identification and localization of disease-specific features at the pixel-level are particularly important for early diagnosis, disease progression monitoring, and effective treatment in medical image analysis. However, conventional diagnostic AI systems lack decision transparency and cannot operate well in environments where there is a lack of pixel-level annotations. In this study, we propose a novel Generative Adversarial Networks (GANs)-based framework, TagGAN, which is tailored for weakly-supervised fine-grained disease map generation from purely image-level labeled data. TagGAN generates a pixel-level disease map during domain translation from an abnormal image to a normal representation. Later, this map is subtracted from the input abnormal image to convert it into its normal counterpart while preserving all the critical anatomical details. Our method is first to generate fine-grained disease maps to visualize disease lesions in a weekly supervised setting without requiring pixel-level annotations. This development enhances the interpretability of diagnostic AI by providing precise visualizations of disease-specific regions. It also introduces automated binary mask generation to assist radiologists. Empirical evaluations carried out on the benchmark datasets, CheXpert, TBX11K, and COVID-19, demonstrate the capability of TagGAN to outperform current top models in accurately identifying disease-specific pixels. This outcome highlights the capability of the proposed model to tag medical images, significantly reducing the workload for radiologists by eliminating the need for binary masks during training.

Problem

• Generates pixel-level disease maps
• Enhances AI diagnostic interpretability
• Automates binary mask generation for radiologists

Innovation

• Generative Adversarial Networks
• weakly-supervised learning
• pixel-level disease map