A picture operation method based on a knowledge graph

Abstract

The invention discloses a picture operation method based on a knowledge graph. The picture operation method is picture search operation and comprises the steps of receiving keywords, 1, obtaining pictures labeled with search words, identifying the labeled search words of the pictures, and if the search words are any one of the keywords, synonyms, hypernyms, subordinate words and related words, returning to the pictures; 2, in the returned picture, the marked search words are in front of a picture row of keywords and keyword synonyms, the marked search words are behind a picture row of keywordhyponyms, keyword hyponyms and keyword related words, and if a user finds a target picture from the arranged picture, the search is completed; If the user does not find the target picture from the arranged pictures, executing the step 3; And step 3, after the associated keywords and the returned picture are labeled with the search words, repeating the step 1 and the step 2 until the target pictureis found in the returned picture. The method can improve the picture search efficiency with small calculation overhead.

Description

technical field [0001] The invention relates to the field of picture manipulation of artificial intelligence, in particular to a picture manipulation method based on knowledge graph. Background technique [0002] In recent years, the number of pictures in smart terminals such as mobile phones has increased year by year, and people are accustomed to using pictures to record every bit of life. It is convenient and quick to record information in the form of pictures, just press the shutter button. But at the same time, when the user takes a large number of pictures, it becomes more and more difficult to find the target picture. [0003] At present, the picture management system of the mobile phone mainly has two operations, namely, searching and browsing. For search operations, there are mainly two existing methods: manual annotation and automatic annotation. Among them, manual labeling is not user-friendly and requires a lot of time; while automatic labeling is limited by t...

AI Technical Summary

Problems solved by technology

Of course, there are also many complex deep learning models that are quite accurate in identifying a certain feature of the image itself, but the computing overhead of most deep learning models is beyond the reach of mobile phones.

For browsing operations, most mobile phones divide pictures into several categories. For example, the built-in photo album of the IOS system divides pictures into people, locations, selfies, food, and screenshots. Browsing books does not match user intent

Images