Gesture recognition method and device

Abstract

The invention discloses a gesture recognition method and device, and relates to the technical field of computer vision. The method comprises the steps that a to-be-processed image is analyzed, a to-be-recognized gesture of a human hand in the to-be-processed image and finger joint information corresponding to the to-be-recognized gesture are obtained, and the to-be-processed image is an image obtained through a three-dimensional depth camera; angle information and distance information in the knuckle information are extracted, the angle information represents the angle between knuckles of the human hand, and the distance information represents the distance between the knuckles of the human hand and the palm root; and the to-be-recognized gesture is identified according to the angle information and the distance information, and thus determining a gesture action. The to-be-processed image is obtained through the three-dimensional depth camera, and the accuracy of the depth information of the to-be-processed image is improved; the features of the knuckle are reflected from multiple dimensions, and the gesture to be recognized is recognized according to the angle information and the distance information, so that the obtained gesture action is more accurate, and the requirement of a user for gesture recognition can be met without consuming excessive computing resources.

Description

technical field [0001] The present application relates to the technical field of computer vision, in particular to a gesture recognition method and device. Background technique [0002] Three-dimensional gesture recognition in computer vision can not only be applied to augmented reality (Augmented Reality, AR) interaction, industrial equipment control, industrial robot control, etc. in the industrial field; three-dimensional gesture recognition can also be applied to consumer-grade AR, consumer-grade somatosensory entertainment and Scenarios for information input such as gesture operations on mobile phones. Through three-dimensional gesture recognition, users can interact with the device naturally and intuitively. The device actively captures the user's gestures, and recognizes and processes the gestures. For example, industrial AR users or industrial virtual reality (Virtual Reality, VR) users can interact with industrial equipment through gestures without using handles or...

AI Technical Summary

Problems solved by technology

[0003] At present, camera gesture recognition can be used to realize three-dimensional gesture recognition. Camera gesture recognition includes monocular cameras and multi-eye cameras; among them, monocular color system (Red Green Blue, RGB) cameras are easier to integrate, but based on monocular RGB cameras It is difficult to directly recognize 3D gestures with traditional gesture recognition methods. Usually, it is necessary to cooperate with key point regression and real-time gesture solutions to complete the 3D gesture recognition of users. poor

[0004] However, the acquisition equipment in the multi-camera implementation is expensive, and it is difficult to directly extract key features from 3D data, which makes this implementation consume a lot of computing resources and cannot support users' needs for low latency and flexible services. deployment requirements

Images