Unbiased Teacher for Semi-Supervised Object Detection

By Yen-Cheng Liu, Chih-Yao Ma, Zijian He, Chia-Wen Kuo, Kan Chen, Peizhao Zhang, Bichen Wu, Zsolt Kira, Péter Vajda. 2021

634 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Memory Replay with Data Compression for Continual Learning (2022, positive): “…of representative old training samples (referred to as memory replay ) can generally achieve the best performance in class-incremental learning (Liu et al., 2021a; Hu et al., 2021) and in numerous other continual learning scenarios, such as audio tasks (Ehret et al., 2020), few-shot (Tao et…”
Semi-Supervised and Unsupervised Deep Visual Learning: A Survey (2022, methodology): “Consistency regularization (Section II-B1) is introduced for semi-supervised object detection to propagate the soft label and bounding boxes assignment on unlabeled images based on dual consistency constraints on clas-siﬁcation and regression [29], [169], [196], [197], [198], [199].”
SODA10M: Towards Large-Scale Object Detection Benchmark for Autonomous Driving (2021, background): “com/google-research/ssl_detection Unbiased Teacher [39] https://github.”
Cross-Domain Adaptive Teacher for Object Detection (2021, methodology): “We also apply weak augmentation (only strong augmentation in Student model) and feed images from the target domain into the Teacher model, which we refer to as “Weak-Strong augmentation”, following Unbiased Teacher (UT) [22].”
Omni-DETR: Omni-Supervised Object Detection with Transformers (2022, positive): “Motivated by the recent successes of student-teacher frameworks for semi-supervised learning [40] and SSOD [30,45], our Omni-DETR is also composed of a student detection network F(x; θs) and a teacher detection network”
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation (2022, methodology): “In this work, we focus on the weak-to-strong consistency regularization framework, that is popularized by FixMatch [55] from the field of semisupervised classification, and then impacts many other relevant tasks [41, 57, 66, 45, 63, 67].”
3rd Continual Learning Workshop Challenge on Egocentric Category and Instance Level Object Understanding (2022, background): “Research does not only cover the advancement of performance under classic learning paradigms (supervised, offline training) [20, 21, 22, 23, 24, 25, 26, 27], but also encompasses areas such as weakly-supervised learning [28, 29, 30, 31], few-shot learning [32, 33, 34, 35, 36, 37, 38], and semi-super”
A Metaverse: Taxonomy, Components, Applications, and Open Challenges (2022, background): “[252] proposed an Unbiased Teacher to identify the pseudo-label bias problem of SS-OD (Semi-Supervised Object Detection).”
Detecting Twenty-thousand Classes using Image-level Supervision (2022, methodology): “Most existing weakly-supervised detection techniques [12, 20, 32, 53, 61] use the weakly labeled data to supervise both the localization and classification subproblems of detection.”