Efficient Image Based Searching for Improving User Search Image Goals

( Vol-2,Issue-4,April 2016 )

Author(s): Manisha Nath, Mr. N Praveen

Total View : 1447
Downloads : 164
Page No: 108-112
DOI:

Keywords:

click-through logs, goal images, image-search goals, semi-supervised clustering, spectral clustering.

Abstract:

The analysis of a user search goals for a query can be very useful in improving search engine relevance and the user experience. Although the research on inferring by user goals and intents for text search has received much attention, so small has been proposed for image search. In this paper, we propose to leverage click session information, which will indicate by high correlations among the clicked images in a session in a user click-through logs, and combine it with the clicked image visual information for inferring the user image-search goals. Since the click session information can serve as past users’ implicit guidance for the clustering the images, more precise user search goals can be obtained. The two strategies are proposed because of combine image visual information for the click session information. Furthermore a classification risk based on approach is also proposed for automatically selecting the optimal number of search goals for a query. Experimental results based on the popular commercial search engine for demonstrate the effectiveness of the proposed method.

Cite This Article:

Share: