Gesture recognition method taking face as reference

Abstract

The invention discloses a gesture recognition method taking face as reference, and belongs to the technical field of gesture recognition. The method includes the following steps: obtaining video data flow and obtaining skeleton point information through processing; according to the skeleton point information, determining palm position information and hand length information of hand length of a user; judging whether the height of the palm of the user away from the ground is larger than a preset height threshold value according to the palm position information: if no, exiting; obtaining the face central position in the video data flow, and judging the distance between the palm position and the face central position is in a standard distance range according to the palm position information: if no, exiting; obtaining an image of a palm area through judgment and performing segmentation, cutting and preprocessing, thereby obtaining a hand mask; and recognizing a fingertip area and then recognizing a gesture. The beneficial effects of the abovementioned technical scheme are that back influence is eliminated, and the circumstances that some invalid gestures are mistaken for gesture instruction input of a user are avoided, thereby improving the accuracy of gesture recognition.

Description

technical field [0001] The present invention relates to the technical field of gesture recognition, in particular to a gesture recognition method using a human face as a reference. Background technique [0002] At present, in the field of smart devices, there are few gesture-controlled smart device methods that can support a large number of gestures. Operate in a gesture recognition environment (for example, in a dark room); 2) The existing gesture recognition method has low accuracy for gesture recognition, and the probability of misrecognition is relatively high; 3) The existing gesture recognition method cannot be accurate. Determine whether the user attempts to perform gesture control. For example, it is impossible to correctly judge whether the purpose of the user's gesture action is to perform gesture control, or just gesture unconsciously. Due to the above reasons, the existing gesture recognition technology cannot be stably applied to smart devices as a control met...

AI Technical Summary

Problems solved by technology

[0002] At present, in the field of smart devices, there are few gesture-controlled smart device methods that can support a large number of gestures. Operate in a gesture recognition environment (for example, in a dark room); 2) The existing gesture recognition method has low accuracy for gesture recognition, and the probability of misrecognition is relatively high; 3) The existing gesture recognition method cannot be accurate. Determine whether the user is trying to perform gesture control

For example, it is impossible to correctly determine whether the purpose of the user's gesture is to control the gesture, or just unconsciously gesture

The above reasons lead to the fact that the existing gesture recognition technology cannot be stably applied to smart devices as a control method

Images