Source data-free domain adaptation for a faster R-CNN. (April 2022)

• Record Type: Journal Article
• Authors: Xiong, Lin; Ye, Mao; Zhang, Dan; Gan, Yan; Liu, Yiguang
• Is Part Of: Pattern recognition. Volume 124(2022)
• Publication Date: 2022-04
• Subjects: Source data-free -- Object detection -- Domain adaptation -- Transfer learning; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2021.108436
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: We proposed a novel prototype-based source data-free domain adaptation method without accessing the source datasets. An iteratively updated scheme for global class prototype was proposed to save the category semantic information. Combining the semantic information of prototype and image features, a more accurate pseudo-labeling method was proposed. Abstract: The existing domain adaptive object detection methods often need to carry a large number of source domain samples for domain adaptation, which is not realistic due to GPU limitations, privacy and physical memory in practical applications. To solve this problem, we propose a source data-free domain adaptive object detection method. Only unlabeled target domain data is used to optimize the source domain model so that it can work better in the target domain. Our method takes Faster R-CNN as baseline. Specifically, we first construct global class prototypes which will be updated in batch iteratively. Then based on the global class prototypes, more accurate pseudo-labels are generated for training the target model. In this way, the source and target domains are also implicitly aligned. Our contributions are 1) a prototype guided domain adaptation method which uses prototypes to mine the semantic category information without accessing the source dataset; 2) a scheme of iteratively updating global class prototype which can handle the class and sample imbalances in the training procedure and 3) a more accurate pseudo-label generation method combining semantic information and image information. On multiple public domain adaptive scenarios, our method achieves the state-of-the-art results in terms of accuracy compared with the Faster R-CNN model and some domain adaptive methods with source datasets.