Method and device for image search

Abstract

The invention relates to a method and a device for image search, wherein the method comprises an offline excavation stage and an online search stage. In the offline excavation stage, aiming at each query in a search behavior log, the method comprises the following steps: S11, collecting the current query and the query related to the current query to form a search term assembly of the current query, S12, unifying the query expressing same semanteme in the current query search term assembly into one interest point to gain various interest points of the current query, S13, storing the interest points of the current query into an interest point database; and in the online search stage, the method comprises the following steps: S21, inquiring the interest point database and determining the interest point of the query input by a user currently, S22, acquiring the image search result of the interest point of the query input by the user currently, and showing the interest points and the image search results of the interest points in the search result page of the query users input in current. The method and device for image search have the advantages that image search requirements of the users can be satisfied more quickly; and the network resource is saved.

Description

【Technical field】 [0001] The invention relates to the technical field of information search, in particular to a method and device for image search. 【Background technique】 [0002] With the continuous development of computer and network technology, search engines have gradually become an important means for people to obtain information, and image search, as one of them, has been more and more widely used. For example, when a user inputs a certain search term (query) in the search box of an image search engine, the image search engine searches for images containing the keyword in the surrounding text of the image, and then sorts the found images and returns them to the user. [0003] In the prior art, when a picture search engine returns picture search results, it sorts the retrieved pictures according to their relevance to a large number of results. Users need to check whether they are the pictures they need one by one from the search results. For example, when the user enter...

AI Technical Summary

Problems solved by technology

[0004] On the one hand, users check whether the pictures are what they want one by one from the massive picture search results, and often need to turn pages multiple times, which wastes both user time and network resources

[0005] On the other hand, when users search, they often need to figure out the query to be entered in order to describe their needs as accurately as possible. However, in many cases, users cannot input accurately at one time, but view the returned pictures after entering the query once. If the search results do not meet your needs, enter a query again by extending the query or transforming the query, etc., and then view the returned image search results. This process may be performed multiple times, which also wastes user time and network resources.

Images