Gesture recognition method and device

Abstract

The invention provides a gesture recognition method and device, and the method and device can achieve the real-time interaction between a deaf-mute and a common person, and achieve the collection of the hand movement posture information of a deaf-mute user through an electromyographic signal sensor, a posture signal sensor and an angle sensor in a flexible glove; via the fusion analysis of the collected data, the method synchronously analyzes data into a gesture frame diagram for output, carries out matching query on the gesture frame diagram and the established word diagram matching database to obtain a voice keyword corresponding to the gesture of the deaf-mute, and outputs a voice expression after the voice keyword is processed by a voice module; and in addition, the spoken language signals of the ordinary person are processed to obtain a gesture posture frame diagram corresponding to the voice of the ordinary person, so that real-time interaction between the deaf-mute and the ordinary person is realized.

Description

technical field [0001] The invention relates to the technical field of gesture recognition, and in particular designs a gesture recognition method and device. Background technique [0002] Since there are many movements of the human hand and the gestures are complex, different movements will also have overlapping parts, making gesture recognition one of the current research problems. Among the complex gestures, sign language is the main language for the deaf-mute to communicate. Sign language is an extremely complex language system and relies on hand movements as the main way for the deaf-mute to communicate with other people. Because sign language is extremely difficult to understand, it is difficult for normal people to accurately recognize the gestures of deaf-mute people unless they are professionals. [0003] At present, the existing sign language recognition methods can be divided into: methods based on data gloves, methods based on myoelectricity and methods based on...

AI Technical Summary

Problems solved by technology

The method based on the data glove needs to wear the data glove, and the data glove can detect the movement information of the hand, including the joint deformation of the fingers, the angle between the fingers, etc. This method directly detects the deformation information of the hand and can better recognize sign language movements. However, in the signals collected by the data gloves, there is no obvious distinction between different gesture signals, and sign language words cannot be effectively segmented. Usually, it is necessary to use externally collected image information for manual segmentation, and the training process is usually time-consuming and laborious.

Although the gesture recognition method based on computer vision is relatively mature, it relies on image processing, and its external factors are also relatively large.

Images