Exploring object relation in mean teacher for cross-domain detection
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 doma...
Saved in:
Main Authors: | , , , , , |
---|---|
Format: | text |
Language: | English |
Published: |
Institutional Knowledge at Singapore Management University
2019
|
Subjects: | |
Online Access: | https://ink.library.smu.edu.sg/sis_research/6457 https://ink.library.smu.edu.sg/context/sis_research/article/7460/viewcontent/Cai_Exploring_Object_Relation_in_Mean_Teacher_for_Cross_Domain_Detection_CVPR_2019_paper.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Singapore Management University |
Language: | English |
id |
sg-smu-ink.sis_research-7460 |
---|---|
record_format |
dspace |
spelling |
sg-smu-ink.sis_research-74602022-01-10T06:10:43Z Exploring object relation in mean teacher for cross-domain detection CAI, Qi PAN, Yingwei NGO, Chong-wah TIAN, Xinmei DUAN, Lingyu YAO, Ting 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 crossdomain 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. 2019-06-01T07:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/6457 info:doi/10.1109/CVPR.2019.01172 https://ink.library.smu.edu.sg/context/sis_research/article/7460/viewcontent/Cai_Exploring_Object_Relation_in_Mean_Teacher_for_Cross_Domain_Detection_CVPR_2019_paper.pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Categorization Recognition: Detection Retrieval Graphics and Human Computer Interfaces OS and Networks |
institution |
Singapore Management University |
building |
SMU Libraries |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
SMU Libraries |
collection |
InK@SMU |
language |
English |
topic |
Categorization Recognition: Detection Retrieval Graphics and Human Computer Interfaces OS and Networks |
spellingShingle |
Categorization Recognition: Detection Retrieval Graphics and Human Computer Interfaces OS and Networks CAI, Qi PAN, Yingwei NGO, Chong-wah TIAN, Xinmei DUAN, Lingyu YAO, Ting Exploring object relation in mean teacher for cross-domain detection |
description |
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 crossdomain 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. |
format |
text |
author |
CAI, Qi PAN, Yingwei NGO, Chong-wah TIAN, Xinmei DUAN, Lingyu YAO, Ting |
author_facet |
CAI, Qi PAN, Yingwei NGO, Chong-wah TIAN, Xinmei DUAN, Lingyu YAO, Ting |
author_sort |
CAI, Qi |
title |
Exploring object relation in mean teacher for cross-domain detection |
title_short |
Exploring object relation in mean teacher for cross-domain detection |
title_full |
Exploring object relation in mean teacher for cross-domain detection |
title_fullStr |
Exploring object relation in mean teacher for cross-domain detection |
title_full_unstemmed |
Exploring object relation in mean teacher for cross-domain detection |
title_sort |
exploring object relation in mean teacher for cross-domain detection |
publisher |
Institutional Knowledge at Singapore Management University |
publishDate |
2019 |
url |
https://ink.library.smu.edu.sg/sis_research/6457 https://ink.library.smu.edu.sg/context/sis_research/article/7460/viewcontent/Cai_Exploring_Object_Relation_in_Mean_Teacher_for_Cross_Domain_Detection_CVPR_2019_paper.pdf |
_version_ |
1770575964044001280 |