arrow
返回

Learning From Box Annotations for Referring Image Segmentation

delete2024-03-01
delete3
PRE
AI
G
Guang Feng
张立鹤 封面图
张立鹤 (Lihe Zhang) *
Z
Zhiwei Hu
卢
卢湖川 (Huchuan Lu)
DOI:10.1109/TNNLS.2022.3201372delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Referring image segmentation (RIS) has obtained an impressive achievement by fully convolutional networks (FCNs). However, previous RIS methods require a large number of pixel-level annotations. In this article, we present a weakly supervised RIS method by using bounding box (BB) annotations. In the first stage, we introduce an adversarial boundary loss to extract the object contour from the BB, which is then used to select appropriate region proposals for pseudoground-truth (PGT) generation. In the second stage, we design a co-training (Co-T) strategy to purify the pseudolabels. Specifically, we train two networks and interactively guide them to pick clean labels for each other's networks, which can weaken the effect of noisy labels on model training. Experiment results on four benchmark datasets demonstrate that the proposed method can produce high-quality masks with a speed of 63 frames/s.
Keyword:
Proposals
Annotations
Image segmentation
Visualization
Semantics
Training
Noise measurement
Adversarial boundary loss
bounding box (BB) annotation
co-training (Co-T) strategy
weakly supervised referring image segmentation (RIS)

期刊

IEEE Transactions on Neural Networks and Learning Systems 封面图
IEEE Transactions on Neural Networks and Learning Systems
IF:
8.9
论文数:
7.6K
被引数:
7.2W

机构

D
Dalian University of Technology
学者数:
6.0W
论文数: 4.4W
被引数: 5.5W
引用论文

引用论文

Sustainable use of sugarcane bagasse ash in cement-based materials
err2019-06-01
err0
PREAI
errPryscila Vinco Andreão; Ahmed R Suleiman; Guilherme Chagas Cordeiro; Moncef L Nehdi
err分享
err收藏
学者 查看更多内容