Identification method and system for social relations of characters in images based on graph reasoning model

Abstract

The invention discloses a method for identifying the social relationship of a person in an image based on a graph inference model, which comprises the following steps: step S1, modeling the coexistence relationship between the social relationship and a semantic object to form a priori knowledge map; S2, combine high-level knowledge map and depth learning to form a graph reasoning network capable of end-to-end learn, and utilizing that propagation mechanism of the graph reasoning network to propagate and learn the information of nodes in the prior knowledge map to obtain weighted graph node output; Step S3, combining the output of the weighted graph nodes and outputting the distribution of the social relation labels of the target person pair through the classification network. By adopting the invention, the efficiency of identifying the social relation of the person in the image can be improved.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a method and system for identifying social relations of characters in images based on a graph reasoning model. Background technique [0002] Social relations are the basic components of social networks and play a very important role in daily life. In the era of people-oriented and rapid development of artificial intelligence, the intelligent analysis of social relations has important applications in many fields, such as multi-target tracking, character trajectory prediction, group activity analysis, human-computer interaction and so on. [0003] There is a certain difficulty in the task of social relationship recognition, because in different scenes, the social relationship of characters may change completely. For example, if the target person pair is in the studio, they are more likely to be a colleague relationship, and if the target person pair is in the kitchen, they...

AI Technical Summary

Problems solved by technology

However, the current methods for solving the task of social relationship recognition either focus on the appearance of the target person pair, or only extract candidate areas that may have relevant information around the target person to assist in the identification of social relations, while ignoring the target person's location. The semantic information of the object cannot accurately identify the social relationship between the characters in the image

Images