Mining User Queries for Image Search using Click-through Approach

The analysis of user serach for a query can be useful in improving search engine performance. Although the text search has received much importance but a little attention has been proposed for image search. In this paper, we propose influencing advantage of click session information, Clickthrough logs maintain clicked images information. The visual information of clicked images is used to infer user-image search goals. The Click session information can be used as past users’ implicit guidance for clustering the images, more precise user search goals can be obtained. “Classification” based approach is proposed for autoclassification of user image search. Experimental results demonstrate the effectiveness of the proposed method.