Adaptation of Weakly Supervised Localization in Histopathology by Debiasing Predictions

2026-03-12

AI Summary

This work addresses the performance degradation of weakly supervised histopathology image localization models under distribution shifts when deployed across organs or institutions, particularly in source-free unsupervised domain adaptation where prediction biases are prone to iterative amplification, impairing both classification and localization. To mitigate this, we propose SFDA-DeP, the first approach to incorporate machine unlearning into this setting. Our method employs entropy-based uncertainty screening to identify over-predicted classes and dynamically suppresses high-confidence biased samples, while jointly optimizing a pixel-level classifier to recover discriminative localization features. This effectively prevents decision boundary drift and alleviates bias accumulation during self-training. Evaluated on multi-center and cross-organ datasets—including GLAS, CAMELYON-16, and CAMELYON-17—SFDA-DeP significantly outperforms existing source-free domain adaptation methods, simultaneously improving classification accuracy and localization precision.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Adversarial Attacks & Robustness
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Weakly Supervised Object Localization (WSOL) models enable joint classification and region-of-interest localization in histology images using only image-class supervision. When deployed in a target domain, distributions shift remains a major cause of performance degradation, especially when applied on new organs or institutions with different staining protocols and scanner characteristics. Under stronger cross-domain shifts, WSOL predictions can become biased toward dominant classes, producing highly skewed pseudo-label distributions in the target domain. Source-Free (Unsupervised) Domain Adaptation (SFDA) methods are commonly employed to address domain shift. However, because they rely on self-training, the initial bias is reinforced over training iterations, degrading both classification and localization tasks. We identify this amplification of prediction bias as a primary obstacle to the SFDA of WSOL models in histopathology. This paper introduces \sfdadep, a method inspired by machine unlearning that formulates SFDA as an iterative process of identifying and correcting prediction bias. It periodically identifies target images from over-predicted classes and selectively reduces the predictive confidence for uncertain (high entropy) images, while preserving confident predictions. This process reduces the drift of decision boundaries and bias toward dominant classes. A jointly optimized pixel-level classifier further restores discriminative localization features under distribution shift. Extensive experiments on cross-organ and -center histopathology benchmarks (glas, CAMELYON-16, CAMELYON-17) with several WSOL models show that SFDA-DeP consistently improves classification and localization over state-of-the-art SFDA baselines. {\small Code: \href{https://anonymous.4open.science/r/SFDA-DeP-1797/}{anonymous.4open.science/r/SFDA-DeP-1797/}}

Problem

• Weakly Supervised Object Localization
• Domain Shift
• Prediction Bias
• Source-Free Domain Adaptation
• Histopathology

Innovation

• Weakly Supervised Object Localization
• Source-Free Domain Adaptation
• Prediction Debiasing
• Histopathology
• Machine Unlearning