[论文解读] Gated CRF Loss for Weakly Supervised Semantic Image Segmentation

Anton Obukhov, Stamatios Georgoulis+2 more|arXiv (Cornell University)|Jun 11, 2019

Advanced Neural Network Applications参考文献 62被引用 55

一句话总结

论文提出了一个 Gated CRF 损失用于弱监督分割中的未标记像素，实现灵活的核与边界聚焦学习，在基于点击和基于涂鸦的监督下实现了最先进的结果，而无需高维滤波。

ABSTRACT

State-of-the-art approaches for semantic segmentation rely on deep convolutional neural networks trained on fully annotated datasets, that have been shown to be notoriously expensive to collect, both in terms of time and money. To remedy this situation, weakly supervised methods leverage other forms of supervision that require substantially less annotation effort, but they typically present an inability to predict precise object boundaries due to approximate nature of the supervisory signals in those regions. While great progress has been made in improving the performance, many of these weakly supervised methods are highly tailored to their own specific settings. This raises challenges in reusing algorithms and making steady progress. In this paper, we intentionally avoid such practices when tackling weakly supervised semantic segmentation. In particular, we train standard neural networks with partial cross-entropy loss function for the labeled pixels and our proposed Gated CRF loss for the unlabeled pixels. The Gated CRF loss is designed to deliver several important assets: 1) it enables flexibility in the kernel construction to mask out influence from undesired pixel positions; 2) it offloads learning contextual relations to CNN and concentrates on semantic boundaries; 3) it does not rely on high-dimensional filtering and thus has a simple implementation. Throughout the paper we present the advantages of the loss function, analyze several aspects of weakly supervised training, and show that our `purist' approach achieves state-of-the-art performance for both click-based and scribble-based annotations.

研究动机与目标

• 通过利用弱监督来降低语义分割的标注负担。
• 开发一种在无需完全监督的情况下有效处理未标记像素的损失函数。
• 引入具有灵活核构造且无需高维滤波的 Gated CRF 损失。
• 证明该方法在基于点击和基于涂鸦的标注下能够达到最先进的结果。

提出的方法

• 在已标记像素上使用部分交叉熵来训练标准 CNN。
• 将所提的 Gated CRF 损失应用于未标记像素以编码上下文关系。
• 使用灵活的核构造来屏蔽掉不希望的像素影响。
• 将上下文学习从 CRF 转移到 CNN，使 Gated CRF 专注于语义边界。
• 避免高维滤波以保持实现简单。

实验结果

研究问题

• RQ1 Gated CRF 损失是否能够通过聚焦边界并屏蔽不希望区域来提升弱监督语义分割的效果？
• RQ2 Gated CRF 损失是否能够在不进行复杂滤波的情况下为基于点击和基于涂鸦的监督提供最先进的性能？
• RQ3 在有效性和简单性方面，所提出的损失与完全监督或其他弱监督方法相比如何？
• RQ4 该训练方法是否在不同的弱监督信号（点击、涂鸦）下都具有鲁棒性？

主要发现

• 该方法在基于点击的标注上实现了最先进的性能。
• 该方法在基于涂鸦的标注上实现了最先进的性能。
• Gated CRF 损失实现了核构造的灵活性，并将学习聚焦于边界。
• 该方法不依赖高维滤波，且实现简单。
• 训练框架保持通用性，并未过度针对单一监督设置定制。