Gesture recognition method of small quantity of training samples based on RGB-D (red, green, blue and depth) data structure

Abstract

The invention discloses a gesture recognition method of a small quantity of training samples based on an RGB-D data structure. The gesture recognition method is implemented by a feature extraction unit, a training unit and a recognition unit, wherein the feature extraction unit is used for extracting three-dimensional sparse SIFT (scale invariable feature transform) features in aligned RGB-D image sequences obtained by an RGB-D camera; the training unit is used for learning models by a small quantity of gesture training samples, and the recognition unit is used for recognizing input continuous gestures. The gesture recognition method can be applied to any camera or equipment, such as Kinect of Microsoft, Xtion PRO of ASUS or Leap Motion of the Leap company, which can provide RGB-D data; and a real-time recognition speed can be realized, so that the method can be used for man-machine interaction, sign language interpretation, smart home, game development and virtual reality.

Description

technical field [0001] The invention relates to a gesture recognition method, which can be applied to human-computer interaction, sign language translation, smart home, game development and virtual reality. Background technique [0002] In traditional gesture recognition, gestures are usually collected by a common camera, and then features are extracted from the RGB video stream. In monocular-based gesture recognition, since only RGB images can be provided, a large number of training samples are usually required to achieve better recognition results; in multi-eye vision, due to the need to calibrate multiple cameras and build 3D models, these Both require complex calculations and cannot achieve real-time results. [0003] In recent years, more and more companies have developed RGB-D cameras. The camera is characterized by the ability to provide RGB images and depth images in real time. For example, in 2010, Microsoft released a camera (Kinect) capable of collecting RGB-D ...

AI Technical Summary

Problems solved by technology

Therefore, when there are only a small number of training samples, the difficulty faced by gesture recognition is how to extract effective features from depth information and color information

[0005] However, in existing methods, there is no small sample data based on RGB-D data to predict the performance of gestures.

Images