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

TL;DR

This paper introduces an online, end-to-end open-set semi-supervised object detection framework with a dual competing head, significantly improving accuracy and efficiency in distinguishing in-distribution from out-of-distribution instances.

Contribution

It proposes a semi-supervised outlier filtering method and a threshold-free dual competing OOD head, advancing OSSOD by enhancing performance and enabling online training.

Findings

Achieves state-of-the-art results on OSSOD benchmarks.

More efficient than offline OOD detection methods.

Easily adaptable to various semi-supervised frameworks.

Abstract

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.