With the explosive growth of the number of images,image search has been drawing much interest from the research community.However,there exists the inconsistency between people's search needs and the search results. Towards this end,this pa-per presents a new feedback framework for interactive image retrieval by adaptively displaying the search results,which can obtain better user experience. The diversity of search results is controlled by introducing a zoom factor,which makes users freely zoom in or out on the results according to their intents. The demo is available at https://youtu.be/hYnx3UkVQEA.%随着图像数量的迅猛增长,图像检索已经在计算机视觉领域引起了越来越多的关注.遗憾的是,其中也存在着人们的检索需求和搜索结果不一致的问题.为此,论文基于改进的聚类算法和重排名算法,提出一种基于反馈框架的交互式图像检索系统来自适应地展示检索结果,从而获得更佳的用户体验.检索结果的多样性由一个缩放因子来控制,用户可以根据自己的意图自由地放大或缩小缩放因子,相当于调整了聚类中心数目k.该系统演示上传于https://youtu.be/hYnx3UkVQEA.
展开▼
