Aggregating multi-scale contextual features from multiple stages for semantic image segmentation. Issue 3 (3rd July 2021)

• Record Type: Journal Article
• Authors: Jiang, Dingchao; Qu, Hua; Zhao, Jihong; Zhao, Jianlong; Hsieh, Meng-Yen
• Is Part Of: Connection science. Volume 33:Issue 3(2021)
• Page Start: 605
• Page End: 622
• Publication Date: 2021-07-03
• Subjects: Deep learning -- semantic segmentation -- multi-scale context; Neural computers -- Periodicals; Artificial intelligence -- Periodicals; Cognitive science -- Periodicals; Connectionism -- Periodicals; 006.3
• DOI: 10.1080/09540091.2020.1862059
• Languages: English
• ISSNs: 0954-0091

Abstract

Abstract : Semantic segmentation plays a vital role in image understanding. Recent studies have attempted to achieve precise pixel-level classification by using deep networks that provide hierarchical features. These methods are trying to effectively utilise multi-level features that are extracted from the data and precisely reconstruct some characteristics of objects that are lost in producing high-level features. In this paper, we propose a multi-scale context U-net (MSCU-net) for semantic image segmentation. This network uses a multi-scale context block (MSCB) to aggregate multi-level features and employs the CRF layer to explicitly model the dependencies among pixels. This network significantly outperforms other state-of-the-art methods on both the PASCAL VOC 2012 and Cityscapes datasets.