Exploring Object Relation in Mean Teacher for Cross-Domain Detection

By Qi Cai, Yingwei Pan, Chong-Wah Ngo, Xinmei Tian, Ling-yu Duan, Ting Yao. 2019

348 citations recorded by the source corpus.

Rendering synthetic data (e.g., 3D CAD-rendered images) to generate annotations for learning deep models in vision tasks has attracted increasing attention in recent years. However, simply applying the models learnt on synthetic images may lead to high generalization error on real images due to domain shift. To address this issue, recent progress in cross-domain recognition has featured the Mean Teacher, which directly simulates unsupervised domain adaptation as semi-supervised learning. The domain gap is thus naturally bridged with consistency regularization in a teacher-student scheme. In this work, we advance this Mean Teacher paradigm to be applicable for cross-domain detection. Specifically, we present Mean Teacher with Object Relations (MTOR) that novelly remolds Mean Teacher under the backbone of Faster R-CNN by integrating the object relations into the measure of consistency cost between teacher and student modules. Technically, MTOR firstly learns relational graphs that capture similarities between pairs of regions for teacher and student respectively. The whole architecture is then optimized with three consistency regularizations: 1) region-level consistency to align the region-level predictions between teacher and student, 2) inter-graph consistency for matching the graph structures between teacher and student, and 3) intra-graph consistency to enhance the similarity between regions of same class within the graph of student. Extensive experiments are conducted on the transfers across Cityscapes, Foggy Cityscapes, and SIM10k, and superior results are reported when comparing to state-of-the-art approaches. More remarkably, we obtain a new record of single model: 22.8% of mAP on Syn2Real detection dataset.

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

Incremental Object Detection via Meta-Learning (2020, neutral): “This has found success in semi-supervised learning [50], [51], [52] and domain adaptation [53], [54], [55], [56] settings.”
Harmonizing Transferability and Discriminability for Adapting Object Detectors (2020, methodology): “Note that the backbone of MTOR is ResNet-50, while the others are VGG-16.”
Unbiased Mean Teacher for Cross-domain Object Detection (2020, methodology): “Various approaches have been proposed to address these issues, including instance and image-level adversarial training [6], strong and weak adversarial training [37], graph-based consistency [1], etc.”
Cross-Domain Adaptive Teacher for Object Detection (2021, methodology): “For example, MTOR [1] employs the Mean Teacher (MT) [40] as its pipeline to identify relations using region-level, inter-graph, and intra-graph consistency.”
Creating Something From Nothing: Unsupervised Knowledge Distillation for Cross-Modal Hashing (2020, background): “Recently, many efforts were made to study what the dark knowledge is [41, 40], and/or how to efficiently take advantage of such knowledge [11, 42, 35, 2].”
Knowledge Distillation and Student-Teacher Learning for Visual Intelligence: A Review and New Outlooks (2020, methodology): “added a target-guided generator to produce augmented images instead of stochastic augmentation, as in [229, 231, 232].”
MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation (2022, methodology): “curriculum [9, 13, 14, 109], exploit the increased domainrobustness of Transformers [34, 35, 83, 102], align the domains with contrastive learning [38, 97], use graph matching [4, 49, 50], or adapt multi-resolution inputs [35].”
ST3D: Self-training for Unsupervised Domain Adaptation on 3D Object Detection (2021, methodology): “Other approaches [31, 55, 21, 5] utilize the self-training strategy to generate pseudo labels for unlabeled target domains.”