集成技术2025，Vol.14Issue(1)：25-38,14.DOI:10.12146/j.issn.2095-3135.20231225001
基于领域自适应预训练的黑暗场景下行为识别研究
Domain-Adaptive Pretraining for Action Recognition in the Dark Scenes
摘要
Abstract
The domain gap between dark scenes and the data used by traditional pretrained models leads to suboptimal performance with the conventional pretrain-finetune approach,and pretraining from scratch is costly.To address this issue,a domain-adaptive pretraining method is proposed to improve action recognition performance in the dark environments.The method integrates an external vision enhancement model for de-darkening to introduce critical knowledge for dark scene processing.It also employs a cross-domain self-distillation framework to reduce the domain gap of visual representations between illuminated and dark scenes.Through extensive experiments in various dark environment action recognition settings,the proposed approach can achieve a Top1 accuracy of 97.19%on the dark dataset of fully supervised action recognition.In the source-free domain adaptation on the Daily-DA dataset,the accuracy can be improved to 49.11%.In the multi-source domain adaptation scenario on the Daily-DA dataset,the Top1 accuracy can reach 54.63%.关键词
黑暗场景/行为识别/迁移学习/领域自适应Key words
dark scenes/action recognition/transfer learning/domain adaptation分类
信息技术与安全科学引用本文复制引用
许清林,乔宇,王亚立..基于领域自适应预训练的黑暗场景下行为识别研究[J].集成技术,2025,14(1):25-38,14.基金项目
国家重点研发计划项目(2022ZD0160505) （2022ZD0160505）
国家自然科学基金项目(62272450) This work is supported by National Key Research and Development Program of China(2022ZD0160505),National Natural Science Foundation of China(62272450) （62272450）
