CrossRectify: Leveraging disagreement for semi-supervised object detection | |
Ma CC(马成丞)1,2; Pan XJ(潘兴甲)3; Ye QX(叶齐祥)4; Tang F(唐帆)5; Dong WM(董未名)1; Xu CS(徐常胜)1 | |
刊名 | Pattern recognition |
2022-12 | |
卷号 | 137页码:109280 |
英文摘要 | Semi-supervised object detection has recently achieved substantial progress. As a mainstream solution, the self-labeling-based methods train the detector on both labeled data and unlabeled data with pseudo labels predicted by the detector itself, but their performances are always limited. Through experimen- tal analysis, we reveal the underlying reason is that the detector is misguided by the incorrect pseudo labels predicted by itself (dubbed self-errors). These self-errors can hurt performance even worse than random-errors, and can be neither discerned nor rectified during the self-labeling process. In this paper, we propose an effective detection framework named CrossRectify, to obtain accurate pseudo labels by simultaneously training two detectors with different initial parameters. Specifically, the proposed ap- proach leverages the disagreements between detectors to discern the self-errors and refines the pseudo label quality by the proposed cross-rectifying mechanism. Extensive experiments show that CrossRectify achieves outperforming performances over various detector structures on 2D and 3D detection benchmarks. |
内容类型 | 期刊论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/54587] |
专题 | 自动化研究所_模式识别国家重点实验室 |
通讯作者 | Tang F(唐帆) |
作者单位 | 1.Chinese Academy of Sciences, Institution of Automation, National Lab Pattern Recognition, Beijing 100190, China 2.University of Chinese Academy of Sciences, School of Artificial Intelligence, Beijing, 100049, China 3.Tencent, Youtu Lab, Shanghai 200233, China 4.School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, 101408, China 5.Jilin University, Changchun 130000, China |
推荐引用方式 GB/T 7714 | Ma CC,Pan XJ,Ye QX,et al. CrossRectify: Leveraging disagreement for semi-supervised object detection[J]. Pattern recognition,2022,137:109280. |
APA | Ma CC,Pan XJ,Ye QX,Tang F,Dong WM,&Xu CS.(2022).CrossRectify: Leveraging disagreement for semi-supervised object detection.Pattern recognition,137,109280. |
MLA | Ma CC,et al."CrossRectify: Leveraging disagreement for semi-supervised object detection".Pattern recognition 137(2022):109280. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论