Action recognition method and recognition system thereof

Abstract

The invention discloses an action recognition method and a recognition system thereof, and the method specifically comprises the following steps: collecting at least one action, decomposing the action, and determining a meta-action; preprocessing the meta-actions to remove weak information in the meta-actions; wherein the step of determining the meta-action comprises the steps of establishing a coordinate system according to the original action, dividing the collected action into a plurality of quadrants, determining the quadrant where the new action is located and the movement track of the new action when the new action occurs, and determining the meta-action according to the movement track of the new quadrant and the movement track of the new action. According to the action recognition method and recognition system provided by the invention, the principle of information entropy can be utilized to filter weak information, remove the weak information in the meta-action sequence and integrate the information of the meta-action sequence, so that the problem of excessive classification in recognition is avoided, and the recognition accuracy and efficiency are further improved.

Description

technical field [0001] The present application relates to the field of action recognition, and in particular, relates to an action recognition method and a recognition system thereof. Background technique [0002] In the natural world, people manipulate objects through touch and express their will through gestures. We hope that humans can interact with computers in the same way. Therefore, the research on "multi-dimensional machine vision motion perception" must first study how to collect three-dimensional motions of the human body, then process the collected complex information, identify effective motions, and finally judge effective motions and analyze the human body. intention of. However, in the traditional effective analysis process, it is not possible to quickly and effectively identify actions with too many classifications, so that action analysis errors or prediction errors may occur. Contents of the invention [0003] The purpose of this application is to proce...

AI Technical Summary

Problems solved by technology

However, in the traditional effective analysis process, it is not possible to quickly and effectively identify actions with too many classifications, so there may be errors in action analysis or prediction errors.

Images