MyGO: Make your Goals Obvious, Avoiding Semantic Confusion in Prostate Cancer Lesion Region Segmentation

2025-07-23

AI Summary

To address semantic ambiguity between lesion and non-lesion regions in prostate cancer (PCa) MRI image segmentation, this paper proposes a Pixel Anchor Module (PAM) that enhances lesion-specific semantic representation via global contextual modeling. Integrated with a self-attention-driven Top-k pixel selection strategy, PAM explicitly focuses on ambiguous, hard-to-discriminate regions, while focal loss mitigates severe class imbalance. Evaluated on the PI-CAI dataset, the method achieves 69.73% IoU and 74.32% Dice—surpassing all prior state-of-the-art approaches. The core contribution lies in the first introduction of a pixel-level anchor mechanism for PCa segmentation, enabling fine-grained semantic decoupling and context-aware feature learning. This innovation bridges the gap between local discriminability and global structural consistency, substantially improving segmentation accuracy for clinically subtle lesions.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Early diagnosis and accurate identification of lesion location and progression in prostate cancer (PCa) are critical for assisting clinicians in formulating effective treatment strategies. However, due to the high semantic homogeneity between lesion and non-lesion areas, existing medical image segmentation methods often struggle to accurately comprehend lesion semantics, resulting in the problem of semantic confusion. To address this challenge, we propose a novel Pixel Anchor Module, which guides the model to discover a sparse set of feature anchors that serve to capture and interpret global contextual information. This mechanism enhances the model's nonlinear representation capacity and improves segmentation accuracy within lesion regions. Moreover, we design a self-attention-based Top_k selection strategy to further refine the identification of these feature anchors, and incorporate a focal loss function to mitigate class imbalance, thereby facilitating more precise semantic interpretation across diverse regions. Our method achieves state-of-the-art performance on the PI-CAI dataset, demonstrating 69.73% IoU and 74.32% Dice scores, and significantly improving prostate cancer lesion detection.

Problem

• Accurate segmentation of prostate cancer lesions due to semantic confusion
• Enhancing model's nonlinear representation for better lesion region identification
• Addressing class imbalance to improve semantic interpretation in diverse regions

Innovation

• Pixel Anchor Module for global context capture
• Self-attention Top_k strategy for anchor refinement
• Focal loss function to address class imbalance