OMG: Observe Multiple Granularities for Natural Language-Based Vehicle Retrieval
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)（2022）
关键词
observe Multiple Granularities,tracked-vehicles,natural language descriptions,smart city construction,given texts,tracked vehicles,surveillance videos,dual-stream framework,text encoder,visual encoder,cross-modal loss function,natural language-based vehicle retrieval task,visual representation,textual representation,objective functions,target features,context features,motion features,global embedding,color-type prompt embedding,semantic features,cross-modal multigranularity contrastive loss function,AI City Challenge Track2
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要