Finger Gesture Recognition Method Based on Depth-of-Field Image

Abstract

The invention discloses a finger gesture recognition method based on a depth-of-field image, which includes the process of turning on a depth-of-field camera to obtain depth-of-field video data; the process of inferring the length of a character's hand and determining the palm position and hand length data; segmenting and cutting out the image of the spherical area of ​​the hand , and the process of preprocessing; and fingertip recognition, and the process of gesture recognition according to its geometric relationship. According to the characteristics of the depth-of-field image, the present invention quickly cuts out the hand area, only analyzes and processes the target area, reduces the computational complexity, and has good adaptability to dynamic scene changes; the fingertip recognition adopts the contour maximum concave point scanning algorithm, The robustness of fingertip recognition is improved. After the fingertips are accurately recognized, each finger is recognized according to the direction vector of the fingers and their geometric relationship, thereby providing recognition of various gestures. The method of the invention is simple, flexible and easy to realize.

Description

Technical field [0001] The invention belongs to the field of human-computer interaction, and specifically relates to a finger gesture recognition method based on a depth image. Background technique [0002] At present, gesture recognition methods at home and abroad are roughly divided into two categories, based on wearable devices and based on traditional vision. Gesture recognition based on wearable devices obtains finger movement characteristic data from sensors such as data gloves and position trackers, and then transfers it to a computer while using neural networks to analyze joint data to obtain gestures to achieve human-computer interaction. The main advantage is that it can measure the posture and gesture of the finger, but it is relatively expensive, which is not conducive to mass promotion and application. The method based on traditional visual recognition uses a common camera to collect gesture video or image information, and then performs recognition processing. Alth...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is that the existing gesture and finger recognition methods in human-computer interaction have the disadvantages of poor or too complicated external lighting conditions, which make the appearance of skin color change greatly and are not credible, require high background complexity and a large amount of calculation In order to solve the problem of low performance, a finger gesture recognition method based on depth-of-field images is provided. According to the characteristics of depth-of-field images, the method uses corresponding algorithms to quickly locate the hands in the image and video streams, identify fingertips, and finally successfully recognize gestures. Thereby improving the flexibility and simplicity of human-computer interaction

Images