arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2305.13802cs.CV
收藏

基于双竞争头的在线开集半监督目标检测

Online Open-set Semi-supervised Object Detection with Dual Competing Head

The University of Tokyo(东京大学);Beijing University of Posts and Telecommunications(北京邮电大学);Tsinghua University(清华大学)

Zerun Wang, Ling Xiao, Liuyu Xiang, Zhaotian Weng, Toshihiko Yamasaki

AI总结:

本文提出一种端到端在线 OSSOD 框架,通过半监督异常值过滤方法和无需阈值的 Dual Competing OOD head 抑制误差累积,在多个基准上实现最先进性能并提升效率。

摘要:

开集半监督目标检测(OSSOD)任务利用包含分布内(ID)和分布外(OOD)实例的实用开集无标签数据集来进行半监督目标检测(SSOD)。OSSOD 中的主要挑战是在伪标签生成过程中区分并过滤 OOD 实例(即异常值),因为 OOD 会影响性能。唯一的 OSSOD 工作采用了一个仅用有标签数据训练的额外离线 OOD 检测网络来解决此问题。然而,有限的有标签数据限制了改进潜力。同时,离线策略导致效率低下。为缓解这些问题,本文提出一种提升性能与效率的端到端在线 OSSOD 框架:1)我们提出一种半监督异常值过滤方法,利用有标签和无标签数据更有效地过滤 OOD 实例。2)我们提出一种无需阈值的 Dual Competing OOD head,通过抑制半监督异常值过滤过程中的误差累积进一步提升性能。3)我们提出的方法是一个在线端到端可训练的 OSSOD 框架。实验结果表明,与现有方法相比,我们的方法在多个 OSSOD 基准上达到了最先进的性能。此外,额外实验表明我们的方法更高效,且能轻松应用于不同的 SSOD 框架以提升其性能。

英文摘要:

Open-set semi-supervised object detection (OSSOD) task leverages practical open-set unlabeled datasets that comprise both in-distribution (ID) and out-of-distribution (OOD) instances for conducting semi-supervised object detection (SSOD). The main challenge in OSSOD is distinguishing and filtering the OOD instances (i.e., outliers) during pseudo-labeling since OODs will affect the performance. The only OSSOD work employs an additional offline OOD detection network trained solely with labeled data to solve this problem. However, the limited labeled data restricts the potential for improvement. Meanwhile, the offline strategy results in low efficiency. To alleviate these issues, this paper proposes an end-to-end online OSSOD framework that improves performance and efficiency: 1) We propose a semi-supervised outlier filtering method that more effectively filters the OOD instances using both labeled and unlabeled data. 2) We propose a threshold-free Dual Competing OOD head that further improves the performance by suppressing the error accumulation during semi-supervised outlier filtering. 3) Our proposed method is an online end-to-end trainable OSSOD framework. Experimental results show that our method achieves state-of-the-art performance on several OSSOD benchmarks compared to existing methods. Moreover, additional experiments show that our method is more efficient and can be easily applied to different SSOD frameworks to boost their performance.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑