Not All Pixels are Equal: Learning Pixel Hardness for Semantic Segmentation

Semantic segmentation has witnessed great progress. Despite the impressive overall results, the segmentation performance in some hard areas (e.g., small objects or thin parts) is still not promising. A straightforward solution is hard sample mining. Yet, most existing hard pixel mining strategies for semantic segmentation often rely on pixel’s loss value, which tends to decrease during training. Intuitively, the pixel hardness for segmentation mainly depends on image structure and is expected to be stable. In this paper, we propose to learn pixel hardness for semantic segmentation by leveraging hardness information contained in global and historical loss values. More precisely, we add a gradient-independent branch for learning a hardness level (HL) map by maximizing hardness-weighted segmentation loss, which is minimized for the segmentation head. This encourages large hardness values in difficult areas, leading to appropriate and stable HL map. Despite its simplicity, the proposed method can be applied to most segmentation methods with no and marginal extra cost during inference and training, respectively. Without bells and whistles, the proposed method achieves consistent improvement (1.37% mIoU on average) over most popular semantic segmentation methods on the Cityscapes dataset, and demonstrates good generalization ability across domains. The source codes are available at this link.