Cargando…

BRefine: Achieving High-Quality Instance Segmentation

Instance segmentation has been developing rapidly in recent years. Mask R-CNN, a two-stage instance segmentation approach, has demonstrated exceptional performance. However, the masks are still very coarse. The downsampling operation of the backbone network and the ROIAlign layer loses much detailed...

Descripción completa

Detalles Bibliográficos
Autores principales: Yu, Jimin, Yang, Xiankun, Zhou, Shangbo, Wang, Shougang, Hu, Shangguo
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9459926/
https://www.ncbi.nlm.nih.gov/pubmed/36080954
http://dx.doi.org/10.3390/s22176499
_version_ 1784786624750551040
author Yu, Jimin
Yang, Xiankun
Zhou, Shangbo
Wang, Shougang
Hu, Shangguo
author_facet Yu, Jimin
Yang, Xiankun
Zhou, Shangbo
Wang, Shougang
Hu, Shangguo
author_sort Yu, Jimin
collection PubMed
description Instance segmentation has been developing rapidly in recent years. Mask R-CNN, a two-stage instance segmentation approach, has demonstrated exceptional performance. However, the masks are still very coarse. The downsampling operation of the backbone network and the ROIAlign layer loses much detailed information, especially from large targets. The sawtooth effect of the edge mask is caused by the lower resolution. A lesser percentage of boundary pixels leads to not-fine segmentation. In this paper, we propose a new method called Boundary Refine (BRefine) that achieves high-quality segmentation. This approach uses FCN as the foundation segmentation architecture, and forms a multistage fusion mask head with multistage fusion detail features to improve mask resolution. However, the FCN architecture causes inconsistencies in multiscale segmentation. BRank and sort loss (BR and S loss) is proposed to solve the problems of segmentation inconsistency and the difficulty of boundary segmentation. It is combined with rank and sort loss, and boundary region loss. BRefine can handle hard-to-partition boundaries and output high-quality masks. On the COCO, LVIS, and Cityscapes datasets, BRefine outperformed Mask R-CNN by 3.0, 4.2, and 3.5 AP, respectively. Furthermore, on the COCO dataset, the large objects improved by 5.0 AP.
format Online
Article
Text
id pubmed-9459926
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-94599262022-09-10 BRefine: Achieving High-Quality Instance Segmentation Yu, Jimin Yang, Xiankun Zhou, Shangbo Wang, Shougang Hu, Shangguo Sensors (Basel) Article Instance segmentation has been developing rapidly in recent years. Mask R-CNN, a two-stage instance segmentation approach, has demonstrated exceptional performance. However, the masks are still very coarse. The downsampling operation of the backbone network and the ROIAlign layer loses much detailed information, especially from large targets. The sawtooth effect of the edge mask is caused by the lower resolution. A lesser percentage of boundary pixels leads to not-fine segmentation. In this paper, we propose a new method called Boundary Refine (BRefine) that achieves high-quality segmentation. This approach uses FCN as the foundation segmentation architecture, and forms a multistage fusion mask head with multistage fusion detail features to improve mask resolution. However, the FCN architecture causes inconsistencies in multiscale segmentation. BRank and sort loss (BR and S loss) is proposed to solve the problems of segmentation inconsistency and the difficulty of boundary segmentation. It is combined with rank and sort loss, and boundary region loss. BRefine can handle hard-to-partition boundaries and output high-quality masks. On the COCO, LVIS, and Cityscapes datasets, BRefine outperformed Mask R-CNN by 3.0, 4.2, and 3.5 AP, respectively. Furthermore, on the COCO dataset, the large objects improved by 5.0 AP. MDPI 2022-08-29 /pmc/articles/PMC9459926/ /pubmed/36080954 http://dx.doi.org/10.3390/s22176499 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Yu, Jimin
Yang, Xiankun
Zhou, Shangbo
Wang, Shougang
Hu, Shangguo
BRefine: Achieving High-Quality Instance Segmentation
title BRefine: Achieving High-Quality Instance Segmentation
title_full BRefine: Achieving High-Quality Instance Segmentation
title_fullStr BRefine: Achieving High-Quality Instance Segmentation
title_full_unstemmed BRefine: Achieving High-Quality Instance Segmentation
title_short BRefine: Achieving High-Quality Instance Segmentation
title_sort brefine: achieving high-quality instance segmentation
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9459926/
https://www.ncbi.nlm.nih.gov/pubmed/36080954
http://dx.doi.org/10.3390/s22176499
work_keys_str_mv AT yujimin brefineachievinghighqualityinstancesegmentation
AT yangxiankun brefineachievinghighqualityinstancesegmentation
AT zhoushangbo brefineachievinghighqualityinstancesegmentation
AT wangshougang brefineachievinghighqualityinstancesegmentation
AT hushangguo brefineachievinghighqualityinstancesegmentation