Unbiased Teacher for Semi-Supervised Object Detection

Semi-supervised learning, i.e., training networks with both labeled and unlabeled data, has made significant progress recently. However, existing works have primarily focused on image classification tasks and neglected object detection which requires more annotation effort. In this work, we revisit the Semi-Supervised Object Detection (SS-OD) and identify the pseudo-labeling bias issue in SS-OD. To address this, we introduce Unbiased Teacher, a simple yet effective approach that jointly trains a student and a gradually progressing teacher in a mutually-beneficial manner. Together with a class-balance loss to downweight overly confident pseudo-labels, Unbiased Teacher consistently improved state-of-the-art methods by significant margins on COCO-standard, COCO-additional, and VOC datasets. Specifically, Unbiased Teacher achieves 6.8 absolute mAP improvements against state-of-the-art method when using 1% of labeled data on MS-COCO, achieves around 10 mAP improvements against the supervised baseline when using only 0.5, 1, 2% of labeled data on MS-COCO.

PDF Abstract ICLR 2021 PDF ICLR 2021 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Semi-Supervised Object Detection COCO 0.5% labeled data Unbiased Teacher mAP 16.94± 0.23 # 4
Semi-Supervised Object Detection COCO 100% labeled data UNBIASED TEACHER mAP 41.3 # 6
Semi-Supervised Object Detection COCO 10% labeled data Unbiased Teacher mAP 31.5± 0.10 # 10
Semi-Supervised Person Bounding Box Detection COCO 1% labeled data Unbiased-Teacher Person Bounding Box AP 39.18 # 2
Semi-Supervised Object Detection COCO 1% labeled data Unbiased Teacher mAP 20.75± 0.12 # 5
Semi-Supervised Object Detection COCO 2% labeled data Unbiased Teacher mAP 24.324.30 ± 0.07 # 11
Semi-Supervised Object Detection COCO 5% labeled data Unbiased Teacher mAP 28.27± 0.11 # 8

Methods


No methods listed for this paper. Add relevant methods here