• Home
  • Search
  • A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data
  • Cite Icon138
  • https://doi.org/10.1609/aaai.v35i10.17029Copy DOI Icon

A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data

  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Unsupervised domain adaptation (UDA) assumes that source and target domain data are freely available and usually trained together to reduce the domain gap. However, considering the data privacy and the inefficiency of data transmission, it is impractical in real scenarios. Hence, it draws our eyes to optimize the network in the target domain without accessing labeled source data. To explore this direction in object detection, for the first time, we propose a source data-free domain adaptive object detection (SFOD) framework via modeling it into a problem of learning with noisy labels. Generally, a straightforward method is to leverage the pre-trained network from the source domain to generate the pseudo labels for target domain optimization. However, it is difficult to evaluate the quality of pseudo labels since no labels are available in target domain. In this paper, self-entropy descent (SED) is a metric proposed to search an appropriate confidence threshold for reliable pseudo label generation without using any handcrafted labels. Nonetheless, completely clean labels are still unattainable. After a thorough experimental analysis, false negatives are found to dominate in the generated noisy labels. Undoubtedly, false negatives mining is helpful for performance improvement, and we ease it to false negatives simulation through data augmentation like Mosaic. Extensive experiments conducted in four representative adaptation tasks have demonstrated that the proposed framework can easily achieve state-of-the-art performance. From another view, it also reminds the UDA community that the labeled source data are not fully exploited in the existing methods.

Similar Papers
  • Research Article
  • Citations16

Overcoming learning bias via Prototypical Feature Compensation for source-free domain adaptation

  • Sep 17, 2024
  • Pattern Recognition
  • Zicheng Pan +3
  • Research Article
  • Citations25

ACT: Semi-supervised Domain-adaptive Medical Image Segmentation with Asymmetric Co-Training.

  • Jan 01, 2022
  • Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
  • Xiaofeng Liu +6
  • Book Chapter
  • Citations8

Domain Adaptation for Eye Segmentation

  • Jan 01, 2020
  • Yiru Shen +2
  • Conference Article
  • Citations11

Combining Dynamic Image and Prediction Ensemble for Cross-Domain Face Anti-Spoofing

  • Jun 06, 2021
  • Lingling Lv +6
  • Research Article

Systematic Analysis of Source-Free Domain Adaptation Methods

  • Mar 27, 2025
  • Frontiers in Computing and Intelligent Systems
  • Zhiyi Miao
  • Research Article

Adaptive integrated weight unsupervised multi-source domain adaptation without source data

  • Apr 22, 2025
  • Complex & Intelligent Systems
  • Zhirui Wang +2
  • PDF
  • Research Article
  • Citations16

Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation.

  • Jun 02, 2022
  • Frontiers in neuroscience
  • Xiaofeng Liu +6
  • Conference Article
  • Citations3

Modeling Sequential Domain Shift through Estimation of Optimal Sub-spaces for Categorization

  • Jan 01, 2014
  • Suranjana Samanta +2
  • Research Article
  • Citations13

Source-free domain adaptation via dynamic pseudo labeling and Self-supervision

  • Jul 22, 2024
  • Pattern Recognition
  • Qiankun Ma +7
  • Research Article
  • Citations121

Divergence-Agnostic Unsupervised Domain Adaptation by Adversarial Attacks.

  • Nov 01, 2022
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Jingjing Li +5
  • Research Article
  • Citations9

Superpixel-guided class-level denoising for unsupervised domain adaptive fundus image segmentation without source data

  • May 26, 2023
  • Computers in Biology and Medicine
  • Meng Zhou +2
  • Research Article

Tight Source Domain Match for Partial Domain Adaptation Based on Maximum Density

  • May 01, 2022
  • Journal of Physics Conference Series
  • Zi-Ao Ni +4
  • Conference Article
  • Citations5

Statistical and Geometrical Alignment using Metric Learning in Domain Adaptation

  • Jul 01, 2020
  • Rakesh Kumar Sanodiya +3
  • Dissertation

Label-efficient visual recognition with transfer learning

  • Jan 01, 2025
  • Jingyi Zhang
  • Research Article
  • Citations37

Online Heterogeneous Transfer Learning by Knowledge Transition

  • May 30, 2019
  • ACM Transactions on Intelligent Systems and Technology
  • Hanrui Wu +5
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.