Current semantic segmentation methods focus only on mining "local" context, i.e., dependencies between pixels within individual images, by context-aggregation modules (e.g., dilated convolution, neural attention) or structure-aware optimization criteria (e.g., IoU-like loss). However, they ignore "global" context of the training data, i.e., rich semantic relations between pixels across different images. Inspired by the recent advance in unsupervised contrastive representation learning, we propose a pixel-wise contrastive framework for semantic segmentation in the fully supervised setting. The core idea is to enforce pixel embeddings belonging to a same semantic class to be more similar than embeddings from different classes. It raises a pixel-wise metric learning paradigm for semantic segmentation, by explicitly exploring the structures of labeled pixels, which were rarely explored before. Our method can be effortlessly incorporated into existing segmentation frameworks without extra overhead during testing. We experimentally show that, with famous segmentation models (i.e., DeepLabV3, HRNet, OCR) and backbones (i.e., ResNet, HR-Net), our method brings consistent performance improvements across diverse datasets (i.e., Cityscapes, PASCAL-Context, COCO-Stuff, CamVid). We expect this work will encourage our community to rethink the current de facto training paradigm in fully supervised semantic segmentation.

翻译：目前的语义分解方法仅注重于开采“ 本地” 环境, 即 个人图像中像素之间的依赖性, 以环境分解模块( 如变异、 神经关注) 或结构分解优化标准( 例如 IoU 类似损失) 的形式 。 然而, 它们忽略了培训数据中的“ 全球” 环境, 即不同图像中像素之间丰富的语义关系。 受最近未受监督的对比代表性学习的进展的启发, 我们提议在完全受监督的环境下为语义分解设置一个像素的对比性框架 。 核心理念是执行属于同一语义类的像素嵌入( 比不同课程中的嵌入更为相似 ) 。 然而, 它们忽略了培训数据分解的“ 全球” 概念, 即不同图像之间的语义关系丰富。 我们的方法可以在测试期间不费力地融入现有的分解框架 。 我们实验显示, 以著名的 HR- 网络 、 REDL 、 RED 和 REV 等 持续 数据 格式 。