Abstract:Object detection methods have achieved remarkable performances when the training and testing data satisfy the assumption of i.i.d. However, the training and testing data may be collected from different domains, and the gap between the domains can significantly degrade the detectors. Test Time Adaptive Object Detection (TTA-OD) is a novel online approach that aims to adapt detectors quickly and make predictions during the testing procedure. TTA-OD is more realistic than the existing unsupervised domain adaptation and source-free unsupervised domain adaptation approaches. For example, self-driving cars need to improve their perception of new environments in the TTA-OD paradigm during driving. To address this, we propose a multi-level feature alignment (MLFA) method for TTA-OD, which is able to adapt the model online based on the steaming target domain data. For a more straightforward adaptation, we select informative foreground and background features from image feature maps and capture their distributions using probabilistic models. Our approach includes: i) global-level feature alignment to align all informative feature distributions, thereby encouraging detectors to extract domain-invariant features, and ii) cluster-level feature alignment to match feature distributions for each category cluster across different domains. Through the multi-level alignment, we can prompt detectors to extract domain-invariant features, as well as align the category-specific components of image features from distinct domains. We conduct extensive experiments to verify the effectiveness of our proposed method. Our code is accessible at https://github.com/yaboliudotug/MLFA.

Intermediate Domain Based Meta Learning Framework for Adaptive Object Detection

Multi-View Domain Adaptive Object Detection on Camera Networks.

Joint Feature-Level And Pixel-Level Domain Adaption For Object Detection In The Wild

Multiscale Domain Adaptive YOLO for Cross-Domain Object Detection

Progressive Domain Adaptation for Object Detection

Meta-UDA: Unsupervised Domain Adaptive Thermal Object Detection using Meta-Learning

Multi-Level Domain Adaptive Learning For Cross-Domain Detection

MLFA: Toward Realistic Test Time Adaptive Object Detection by Multi-Level Feature Alignment

Domain Adaptive Object Detection for Autonomous Driving under Foggy Weather

Meta-SSD: Towards Fast Adaptation for Few-Shot Object Detection with Meta-Learning

Integrated Multiscale Domain Adaptive YOLO

Multilevel Unsupervised Domain Adaptation for Single-Stage Object Detection in Remote Sensing Images

UMS2-ODNet: Unified-scale Domain Adaptation Mechanism Driven Object Detection Network with Multi-Scale Attention

Domain Adaptive Object Detection via Balancing Between Self-Training and Adversarial Learning

Spatial Alignment for Unsupervised Domain Adaptive Single-Stage Object Detection

An Adaptive Detector for Few Shot Object Detection.

Open-Scenario Domain Adaptive Object Detection in Autonomous Driving

Progressive Domain Adaptive Object Detection Based on Self‐Attention in Foggy Weather

Unsupervised Domain Adaptation for Object Detection Through Mixed-Domain and Co-Training Learning

Online Adaptation Through Meta-Learning for Stereo Depth Estimation

Domain Adaptation based Object Detection for Autonomous Driving in Foggy and Rainy Weather