ProCNS: Progressive Prototype Calibration and Noise Suppression for Weakly-Supervised Medical Image Segmentation
Y. Liu, L. Lin, K. K. Y. Wong, X. Tang
TL;DR
ProCNS addresses the core challenge of weakly-supervised medical image segmentation by introducing two synergistic modules: PRSA, which calibrates prototypes through progressive, multi-scale affinities between spatial and semantic elements, and ANPM, which adaptively masks noisy regions to prevent erroneous prototype updates. The framework also provides soft supervision for identified noisy regions and can be integrated as a plug-in with existing WSS methods. Extensive experiments across six diverse medical imaging tasks demonstrate improved Dice scores and competitive efficiency, validating the approach's effectiveness in handling annotation sparsity and boundary ambiguity. The work highlights the potential for prototype-driven refinement and noise-aware supervision to advance label-efficient medical image analysis, while also exploring integration with foundation-model predictions as a future direction.
Abstract
Weakly-supervised segmentation (WSS) has emerged as a solution to mitigate the conflict between annotation cost and model performance by adopting sparse annotation formats (e.g., point, scribble, block, etc.). Typical approaches attempt to exploit anatomy and topology priors to directly expand sparse annotations into pseudo-labels. However, due to a lack of attention to the ambiguous edges in medical images and insufficient exploration of sparse supervision, existing approaches tend to generate erroneous and overconfident pseudo proposals in noisy regions, leading to cumulative model error and performance degradation. In this work, we propose a novel WSS approach, named ProCNS, encompassing two synergistic modules devised with the principles of progressive prototype calibration and noise suppression. Specifically, we design a Prototype-based Regional Spatial Affinity (PRSA) loss to maximize the pair-wise affinities between spatial and semantic elements, providing our model of interest with more reliable guidance. The affinities are derived from the input images and the prototype-refined predictions. Meanwhile, we propose an Adaptive Noise Perception and Masking (ANPM) module to obtain more enriched and representative prototype representations, which adaptively identifies and masks noisy regions within the pseudo proposals, reducing potential erroneous interference during prototype computation. Furthermore, we generate specialized soft pseudo-labels for the noisy regions identified by ANPM, providing supplementary supervision. Extensive experiments on six medical image segmentation tasks involving different modalities demonstrate that the proposed framework significantly outperforms representative state-of-the-art methods.