Portrait clustering method and device and storage medium

Abstract

The invention discloses a portrait clustering method and device and a storage medium, and the method comprises the steps: obtaining an image sequence; extracting a portrait picture in the image and portrait features in the portrait picture, and performing clustering processing to obtain clustering results of different portrait features; according to relevance of different portrait features, clustering results of the different portrait features are fused, and the same kind of portraits, the suspected same kind of portraits and different kinds of portraits are obtained; re-weighting the similarity of different portrait features of the suspected targets in the same kind of suspected portraits to obtain the comprehensive similarity of the suspected targets; and comparing the comprehensive similarity with a preset threshold, and classifying the suspected same type of portraits into the same type of portraits or different types of portraits. According to the method, the suspected targets of the suspected same-type portraits are subjected to secondary verification, so that the suspected same-type portraits are further accurately classified, data which cannot be in a middle section are eliminated, and the recall rate of clustering is effectively increased.

Description

Technical field[0001]The present application belongs to the field of video analysis, and is specifically related to portrait clustering methods, equipment and storage media.Background technique[0002]With the development of related technologies, face clusters have been widely used. For example, mobile photo albums, through face clustering methods, you can get the face images of the same person to a group; if you solve the case in the public security industry, it is necessary to retrieve in the massive portrait library according to the suspect image of the unrecognized identity. Based on this need to establish a single-file information base, the image of the same person in the information library belongs to the same category.[0003]The image can be clustering the image in the existing related technique, but the existing face clustering method is low, so that there is more image data that cannot be classified, resulting in the recall of the clustering method. low.Inventive content[0004]...

AI Technical Summary

Problems solved by technology

[0004] This application provides a portrait clustering method, device and storage medium to solve the technical problem of the low recall rate of the clustering method

Images