Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/151821
Title: Weakly supervised segmentation with maximum bipartite graph matching
Authors: Liu, Weide
Zhang, Chi
Lin, Guosheng
Hung, Tzu-Yi
Miao, Chunyan
Keywords: Engineering::Computer science and engineering
Issue Date: 2020
Source: Liu, W., Zhang, C., Lin, G., Hung, T. & Miao, C. (2020). Weakly supervised segmentation with maximum bipartite graph matching. MM '20: Proceedings of the 28th ACM International Conference on Multimedia, 2085-2094. https://dx.doi.org/10.1145/3394171.3413652
Project: SMA-RP10
AISG-RP-2018-003
RG126/17 (S)
RG28/18 (S)
RG22/19 (S)
Abstract: In the weakly supervised segmentation task with only image-level labels, a common step in many existing algorithms is first to locate the image regions corresponding to each existing class with the Class Activation Maps (CAMs), and then generate the pseudo ground truth masks based on the CAMs to train a segmentation network in the fully supervised manner. The quality of the CAMs has a crucial impact on the performance of the segmentation model. We propose to improve the CAMs from a novel graph perspective. We model paired images containing common classes with a bipartite graph and use the maximum matching algorithm to locate corresponding areas in two images. The matching areas are then used to refine the predicted object regions in the CAMs. The experiments on Pascal VOC 2012 dataset show that our network can effectively boost the performance of the baseline model and achieves new state-of-the-art performance.
URI: https://hdl.handle.net/10356/151821
ISBN: 9781450379885
DOI: 10.1145/3394171.3413652
Rights: © 2020 Association for Computing Machinery. All rights reserved.
Fulltext Permission: none
Fulltext Availability: No Fulltext
Appears in Collections:SCSE Conference Papers

SCOPUSTM   
Citations 20

9
Updated on Dec 3, 2022

Web of ScienceTM
Citations 50

1
Updated on Nov 29, 2022

Page view(s)

156
Updated on Dec 5, 2022

Google ScholarTM

Check

Altmetric


Plumx

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.