发表机构
School of Computer Science, Hangzhou Dianzi University; Hangzhou Institute of Technology, Xidian University; University of Yamanashi; Department of Electronic Engineering, School of Information Science and Engineering, Fudan University; Ruian People’s Hospital; School of Software, Shandong University（杭州电子科技大学计算机学院; 西安电子科技大学杭州研究院; 山梨大学; 复旦大学信息科学与工程学院电子工程系; 瑞安市人民医院; 山东大学软件学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对弱监督体积分割中CAM不可靠问题，提出无训练无原型框架，利用VRAA抑制噪声放大信号，用BER机制校正不合理激活，该方法与模型无关，集成方便，实验显示相比现有方法有显著提升，减少推理时间。
AI 中文摘要
弱监督分割严重依赖类激活映射（CAM）来初步定位目标区域，但CAM往往有噪声且容易出现灾难性故障。现有补救措施通常会增加额外训练阶段或原型学习，增加计算成本并降低鲁棒性。本文提出一个无训练、无原型的框架，通过利用体积数据中的时间和结构相干性来校正不可靠的CAM。该方法基于两个关键组件：一是引入方差降低激活聚合（VRAA）抑制噪声并放大相干语义信号，通过将CAM建模为高维随机向量提供理论依据；二是设计双向极值校正（BER）机制，通过双向极值检查检测并校正不合理激活，有效减轻极值故障且无需学习额外参数。该方法与模型无关，可无缝集成到现有管道。在多个公共基准上的大量实验表明，该方法比现有弱监督方法有显著改进，Dice高达20%的提升，mIoU高达40%的提升，同时推理时间减少5倍以上。这些结果表明，利用相干性作为隐式归纳偏差可产生一种稳定弱监督体积分割的原则性且高效的方法。
英文摘要
Weakly supervised segmentation relies heavily on class activation maps (CAMs) to initially localize target regions. However, CAMs are often noisy and prone to catastrophic failures. Existing remedies typically introduce additional training stages or prototype learning, increasing computational cost and reducing robustness. In this paper, we propose a training-free prototype-free framework that rectifies unreliable CAMs by exploiting temporal and structural coherence in volumetric data as a free lunch. Our approach is built on two key components. First, we introduce Variance-Reduced Activation Aggregation (VRAA) which suppresses noise and amplify coherent semantic signals. We provide a theoretical justification by modeling CAMs as high-dimensional random vectors and show that aggregation yields provable variance reduction. Second, we design a Bidirectional Extremity Rectification (BER) mechanism that detects and rectifies implausible activations through bidirectional extremity checks, effectively mitigating extreme-value failures without learning additional parameters. Our method is model-agnostic and can be seamlessly integrated with existing pipelines. Extensive experiments on multiple public benchmarks demonstrate substantial improvements over state-of-the-art weakly supervised methods, achieving up to 20% Dice and 40% mIoU gains while reducing inference time by more than 5 times. These results indicate that leveraging coherence as an implicit inductive bias yields a principled and efficient approach to stabilizing weakly supervised volumetric segmentation. Our code will be available.
