Method for determining roles of staffs on basis of behaviors

Abstract

The invention discloses a method for determining roles of staffs on the basis of behaviors, comprising the steps of obtaining a video source; extracting tracks of the staffs in the video source and marking the roles of the staffs; extracting the set of positions where the staffs enter and leave a scene in all the tracks as well as the set of positions where the staffs stay in the scene; carrying out clustering calculation on the sets of positions according to an expectation maximization algorithm so as to obtain the positions of an entrance and an exit of the scene and the positions of objects in the scene; defining the space relationship among the entrance and the exit of the scene as well as the objects in the scene according to the positions of the entrance and the exit of the scene as well as the positions of the objects in the scene; defining the staying time of the staffs at the objects in the scene; defining the actions of the staffs according to the space relationship and the time; defining the activities of the staffs according to the actions of the staffs; defining the behaviors of the staffs according to the activities of the staffs; and defining the probability of the roles of the staffs according to the behaviors of the staffs. The method provided by the invention has the advantages of easiness in realization and good effect by only depending on video information and utilizing the easily-detected interaction relationship between human beings and environmental objects.

Description

technical field [0001] The invention relates to the field of video processing, and more specifically, the invention relates to a method for determining a person's role based on behavior. Background technique [0002] With the development of multimedia technology and artificial intelligence technology, automatically extracting useful information from original multimedia information can be used to complete the compression and storage of multimedia materials and text retrieval, and can also drive other data and trigger other behaviors, thereby greatly reducing Human participation expands the scope of application of human-computer interaction. The role of people plays a very important role in real life, so it is of great significance to use artificial intelligence technology to extract the role information of people appearing in the original data from the multimedia data. [0003] At present, the method of determining personnel roles from multimedia data mainly starts with thre...

AI Technical Summary

Problems solved by technology

[0004] However, these methods all have certain defects: because the text information requires a lot of manual operations, the audio has certain spatial limitations (the voice information will not be collected if the distance is too far away), and it is difficult to separate the mixed voice into each individual the sound of

In this way, a lot of information about the interaction between people and the scene is wasted, and because the interaction between people is difficult to detect, this method is not very practical and the effect is relatively poor.

Images