Abstract:Source-Free domain adaptive Object Detection (SFOD) is a promising strategy for deploying trained detectors to new, unlabeled domains without accessing source data, addressing significant concerns around data privacy and efficiency. Most SFOD methods leverage a Mean-Teacher (MT) self-training paradigm relying heavily on High-confidence Pseudo Labels (HPL). However, these HPL often overlook small instances that undergo significant appearance changes with domain shifts. Additionally, HPL ignore instances with low confidence due to the scarcity of training samples, resulting in biased adaptation toward familiar instances from the source domain. To address this limitation, we introduce the Low-confidence Pseudo Label Distillation (LPLD) loss within the Mean-Teacher based SFOD framework. This novel approach is designed to leverage the proposals from Region Proposal Network (RPN), which potentially encompasses hard-to-detect objects in unfamiliar domains. Initially, we extract HPL using a standard pseudo-labeling technique and mine a set of Low-confidence Pseudo Labels (LPL) from proposals generated by RPN, leaving those that do not overlap significantly with HPL. These LPL are further refined by leveraging class-relation information and reducing the effect of inherent noise for the LPLD loss calculation. Furthermore, we use feature distance to adaptively weight the LPLD loss to focus on LPL containing a larger foreground area. Our method outperforms previous SFOD methods on four cross-domain object detection benchmarks. Extensive experiments demonstrate that our LPLD loss leads to effective adaptation by reducing false negatives and facilitating the use of domain-invariant knowledge from the source model. Code is available at <a class="link-external link-https" href="https://github.com/junia3/LPLD" rel="external noopener nofollow">this https URL</a>.

CLIP-Guided Source-Free Object Detection in Aerial Images

Source-free Domain Adaptive Object Detection in Remote Sensing Images

Source-Free Object Detection by Learning to Overlook Domain Style

Simplifying Source-Free Domain Adaptation for Object Detection: Effective Self-Training Strategies and Performance Insights

Adversarial Alignment for Source Free Object Detection

SF-FSDA: Source-Free Few-Shot Domain Adaptive Object Detection with Efficient Labeled Data Factory

Multi-level domain perturbation for source-free object detection in remote sensing images

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

Exploiting Low-confidence Pseudo-labels for Source-free Object Detection

Guiding Clean Features for Object Detection in Remote Sensing Images

Proposal-level reliable feature-guided contrastive learning for SFOD

SSDA3D: Semi-supervised Domain Adaptation for 3D Object Detection from Point Cloud

Enhancing Source-Free Domain Adaptive Object Detection with Low-confidence Pseudo Label Distillation

Refined Pseudo Labeling for Source-Free Domain Adaptive Object Detection

Object Detection for Aerial Images With Feature Enhancement and Soft Label Assignment

Object Detection in Aerial Images in Scarce Data Regimes

Source-free Domain Adaptation for Video Object Detection Under Adverse Image Conditions

Multi-source-free Domain Adaptive Object Detection

Source-Free Open Compound Domain Adaptation in Semantic Segmentation

Toward Open Vocabulary Aerial Object Detection with CLIP-Activated Student-Teacher Learning