基于本体和相关反馈的动漫图片检索研究

Research of Animation Image Retrieval System Based on Ontology and Relevance Feedback

随着动漫产业的迅速发展,传统的手工绘图已逐渐被计算机软件绘图所代替。在动漫创作时,通过检索系统对图片进行共享和复用,可以减少重复劳动和降低开发成本。传统检索系统集中在关键字匹配方面,查准率不高。为解决此问题,提出一种基于本体和相关反馈相结合的动漫图片检索方法,并通过实验验证了系统的有效性和实用性。

With the rapid development of animation industry, the traditional hand-drawing has been gradually replaced by computer graphics software. Establish a cartoon image gallery to share and reuse these images can effectively reduce du- plication of effort and reduce development costs. Since the traditional retrieval system focn^d on the keyword matching, precision is not high. In response to solving the problem, this article presents an animation image retrieval method which is based on the combination of ontology and relevance feedback retrieval technology. Experiments show that retrieval sys- tem is accuracy and practicality.