Joint Semantic Segmentation and Boundary Detection using Iterative Pyramid Contexts

• DOI: 10.48550/arxiv.2004.07684
• Publication Date: 2020-01-01
• Authors: Mingmin Zhen; Jinglu Wang; Lei Zhou; Shiwei Li; Tianwei Shen; Jiaxiang Shang; Tian Fang; Long Quan
• Keywords: Pyramid (geometry)

In this paper, we present a joint multi-task learning framework for semantic segmentation and boundary detection. The critical component in the is iterative pyramid context module (PCM), which couples two tasks stores shared latent semantics to interact between tasks. For detection, propose novel spatial gradient fusion suppress nonsemantic edges. As detection dual task of segmentation, introduce loss function with consistency constraint improve pixel accuracy segmentation. Our extensive experiments demonstrate superior performance over state-of-the-art works, not only but also particular, mean IoU score 81:8% on Cityscapes test set achieved without using coarse data or any external previous works by 9.9% terms AP 6:8% MF(ODS).