EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection
This paper proposes a new method for addressing the problem of unsupervised domain adaptation for robust object detection. To this end, we propose an energy-based curriculum for progressively adapting a model, thereby mitigating the pseudo-label noise caused by domain shifts. Throughout the adaptati...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10192408/ |
| _version_ | 1797746157477167104 |
|---|---|
| author | Amin Banitalebi-Dehkordi Abdollah Amirkhani Alireza Mohammadinasab |
| author_facet | Amin Banitalebi-Dehkordi Abdollah Amirkhani Alireza Mohammadinasab |
| author_sort | Amin Banitalebi-Dehkordi |
| collection | DOAJ |
| description | This paper proposes a new method for addressing the problem of unsupervised domain adaptation for robust object detection. To this end, we propose an energy-based curriculum for progressively adapting a model, thereby mitigating the pseudo-label noise caused by domain shifts. Throughout the adaptation process, we also make use of spatial domain mixing as well as knowledge distillation to improve the pseudo-labels reliability. Our method does not require any modifications in the model architecture or any special training tricks or complications. Our end-to-end pipeline, although simple, proves effective in adapting object detector neural networks. To verify our method, we perform an extensive systematic set of experiments on: synthetic-to-real scenario, cross-camera setup, cross-domain artistic datasets, and image corruption benchmarks, and establish a new state-of-the-art in several cases. For example, compared to the best existing baselines, our Energy-Based Curriculum learning method for robust object Detection (EBCDet), achieves: 1–3 % AP50 improvement on Sim10k-to-Cityscapes and KITTI-to-Cityscapes, 3–4 % AP50 boost on Pascal-VOC-to- Comic, WaterColor, and ClipArt, and 1-5% relative robustness improvement on Pascal-C, COCO-C, and Cityscapes-C (1-2 % absolute mPC). Code is available at: <uri>https://github.com/AutomotiveML/EBCDet</uri>. |
| first_indexed | 2024-03-12T15:33:00Z |
| format | Article |
| id | doaj.art-f806ae0f2cd149dbad3d4d709a7745a8 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-12T15:33:00Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-f806ae0f2cd149dbad3d4d709a7745a82023-08-09T23:01:22ZengIEEEIEEE Access2169-35362023-01-0111778107782510.1109/ACCESS.2023.329836910192408EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object DetectionAmin Banitalebi-Dehkordi0Abdollah Amirkhani1https://orcid.org/0000-0001-6891-4528Alireza Mohammadinasab2Big Data and Intelligence Platform Laboratory, Huawei Technologies Canada Company Ltd., Markham, CanadaSchool of Automotive Engineering, Iran University of Science and Technology, Tehran, IranSchool of Automotive Engineering, Iran University of Science and Technology, Tehran, IranThis paper proposes a new method for addressing the problem of unsupervised domain adaptation for robust object detection. To this end, we propose an energy-based curriculum for progressively adapting a model, thereby mitigating the pseudo-label noise caused by domain shifts. Throughout the adaptation process, we also make use of spatial domain mixing as well as knowledge distillation to improve the pseudo-labels reliability. Our method does not require any modifications in the model architecture or any special training tricks or complications. Our end-to-end pipeline, although simple, proves effective in adapting object detector neural networks. To verify our method, we perform an extensive systematic set of experiments on: synthetic-to-real scenario, cross-camera setup, cross-domain artistic datasets, and image corruption benchmarks, and establish a new state-of-the-art in several cases. For example, compared to the best existing baselines, our Energy-Based Curriculum learning method for robust object Detection (EBCDet), achieves: 1–3 % AP50 improvement on Sim10k-to-Cityscapes and KITTI-to-Cityscapes, 3–4 % AP50 boost on Pascal-VOC-to- Comic, WaterColor, and ClipArt, and 1-5% relative robustness improvement on Pascal-C, COCO-C, and Cityscapes-C (1-2 % absolute mPC). Code is available at: <uri>https://github.com/AutomotiveML/EBCDet</uri>.https://ieeexplore.ieee.org/document/10192408/Object detectiondomain adaptationenergymodel robustnesscurriculum learning |
| spellingShingle | Amin Banitalebi-Dehkordi Abdollah Amirkhani Alireza Mohammadinasab EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection IEEE Access Object detection domain adaptation energy model robustness curriculum learning |
| title | EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection |
| title_full | EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection |
| title_fullStr | EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection |
| title_full_unstemmed | EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection |
| title_short | EBCDet: Energy-Based Curriculum for Robust Domain Adaptive Object Detection |
| title_sort | ebcdet energy based curriculum for robust domain adaptive object detection |
| topic | Object detection domain adaptation energy model robustness curriculum learning |
| url | https://ieeexplore.ieee.org/document/10192408/ |
| work_keys_str_mv | AT aminbanitalebidehkordi ebcdetenergybasedcurriculumforrobustdomainadaptiveobjectdetection AT abdollahamirkhani ebcdetenergybasedcurriculumforrobustdomainadaptiveobjectdetection AT alirezamohammadinasab ebcdetenergybasedcurriculumforrobustdomainadaptiveobjectdetection |