作者: Jingyu Cui , Fang Wen , Xiaoou Tang
DOI: 10.1109/ICME.2010.5583220
关键词:
摘要: We propose three innovative interactive methods to let computer better understand user intention in content-based image retrieval: 1. Smart list induces intention, thereby improves search results by intention-specific schema; 2. Reference strokes interaction allows specify detail about the pointing out interested regions; 3. Natural feedback easily collects data of relevance feedbacks boost performance system. Systematic study shows that proposed mechanism efficiency, reduces workload, and enhances experience.