Related method and device for sorting image object search results

Abstract

Provided in the present invention are a method and device for ranking search results of picture objects. The method for ranking search results of picture objects comprises: in response to receiving a search request of a user, obtaining picture objects matching the search request; extracting visual characteristics of the obtained picture objects; obtaining comprehensive preference scores of the user to the obtained picture objects on the basis of preference scores of the user to the visual characteristics; ranking the picture objects according to the comprehensive preference scores of the user to the obtained picture objects. The present invention enables ranking of the search results of the picture objects to be more accurate and close to the user interest.

Description

technical field [0001] The invention relates to the technical field of the Internet, in particular to a related method and device for sorting image object search results. Background technique [0002] With the development of Internet technology, the amount of information on the Internet is also increasing day by day. When users search for the information they need through search engines, it is often because the search keywords entered by the user cannot accurately describe the required information or match the user search keywords. Due to the large amount of information and other reasons, search engines return a large amount of relevant information as search results. The more information contained in the search results, the more difficult it will be for users to select the information they need in the search results. [0003] In order to solve the above problems, some search engines provide a solution for sorting search results based on user interests, specifically: clusteri...

AI Technical Summary

Problems solved by technology

Among them, if there are many text features in a search result, an information extraction operation is required, because it is difficult to quantify when extracting information, and if the text features are not regular table types, it is difficult to extract the features; at the same time, due to different search results Different text description methods may be used for the same attribute in the results. For example, the attribute silk and real silk of the product are actually the same attribute. Therefore, the search results need to be normalized. If the normalization is not accurate, it will not be possible. Identify the two same attributes to the same point of interest; in addition, because the same keyword represents different meanings under different categories, for example, for the keyword apple, it represents under different categories such as fruit, movie, 3C, etc. The meaning is different, so it is also necessary to calculate its semantic scene based on the context information to disambiguate the keyword. If the context information cannot be obtained, the disambiguation result will be affected. Then the search result obtained by searching based on the disambiguated keyword The result will contain a large amount of information that does not match the real needs of users, and clustering based on this information will make it difficult to accurately express the real points of interest of users

[0004] It can be seen that determining the information that users are interested in based on text features requires complex technical processing such as information extraction, normalization, and disambiguation of text features. The accuracy of the extracted attribute information for user interest analysis is low, and it is difficult to accurately express the real interest points of users. After sorting the search results based on this, it is also difficult for users to find the information they really need

Images