Multi-Prototype Guided Source-Free Domain Adaptive Object Detection for Autonomous Driving
- 30 November 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (1) , 1589-1601
Abstract
Source-free domain adaptive object detection (source-free DAOD) seeks to adapt a detector pre-trained on a source domain to an unlabeled target domain without requiring access to annotated source domain data. To address challenges posed by domain shifts, current source-free DAOD approaches mainly rely on the self-training paradigm, where pseudo labels are predicted and employed to fine-tune the detector on unlabeled target domain. However, these methods often encounter issues related to intra-class variation, resulting in category-specific biases and noisy pseudo labels. In response, we present an effective Multi-Prototype Guided source-free DAOD method, dubbed MPG, consisting of two key components: multi-prototype guided pseudo labeling (MPPL) and multi-prototype guided consistency regularization (MPCR) modules. In the MPPL module, we construct category-specific multiple prototypes to better represent the category with intra-class variations. Specifically, multiple prototypes with adaptive cluster centroids are introduced for each category to effectively capture the intra-class variations. Through the implementation of the proposed MPPL module, we derive more accurate pseudo labels by assessing the proximity of instance features to multiple category prototypes. In the MPCR module, we introduce multi-level consistency regularization, including prototype-based consistency and prediction consistency, which encourages the model to overlook style perturbations and learn domain-invariant representations. Extensive experiments on five public driving datasets demonstrate that MPG outperforms existing state-of-the-art methods, showcasing its effectiveness in adapting object detectors to target domains.Keywords
Funding Information
- National Key Research and Development Plan of China (2020AAA0108902)
- National Natural Science Foundation of China (62206288)
This publication has 55 references indexed in Scilit:
- Frequency-based pseudo-domain generation for domain generalizable object detectionNeurocomputing, 2023
- Self-training transformer for source-free domain adaptationApplied Intelligence, 2022
- Multi-adversarial Faster-RCNN with Paradigm Teacher for Unrestricted Object DetectionInternational Journal of Computer Vision, 2022
- BMD: A General Class-Balanced Multicentric Dynamic Prototype Strategy for Source-Free Domain Adaptation2022
- Source free domain adaptation for medical image segmentation with fourier style miningMedical Image Analysis, 2022
- Source data-free domain adaptation for a faster R-CNNPattern Recognition, 2022
- Domain Adaptive Faster R-CNN for Object Detection in the Wild2018
- Semantic Foggy Scene Understanding with Synthetic DataInternational Journal of Computer Vision, 2018
- The Cityscapes Dataset for Semantic Urban Scene Understanding2016
- Large-Scale Machine Learning with Stochastic Gradient Descent2010