Gesture recognition method based on extreme learning machine

Abstract

The invention discloses a gesture recognition method based on an extreme learning machine. The gesture recognition method comprises the steps of acquiring an RGB image by using an ordinary camera, and acquiring an infrared image by using an infrared camera; determining a pixel region belonging to a human body in the infrared image by adopting a method of threshold segmentation, then acquiring a human body RGB image by combining the RGB image, acquiring a skin color region in the human body RGB image via a skin color model according to the human body RGB image, then judging the shape complexity of the skin region by using a threshold method, and acquiring a gesture binary image; extracting an HOG features and an Hu moments of a gesture region to act as feature vectors; and finally applying the extreme learning machine to classification for the gesture feature vectors so as to completing a gesture recognition task. The gesture recognition accuracy is improved through combining the HOG features and the Hu moment features. Meanwhile, the training speed and the recognition speed of gesture recognition are effectively improved by applying the extreme learning machine, and a gesture recognition system is enabled to have the advantages of rapidness, high accuracy, high generalization and the like.

Description

technical field [0001] The invention relates to the field of image recognition, in particular to a gesture recognition method based on an extreme learning machine. Background technique [0002] Today, the main means of human-computer interaction is still the mouse and keyboard. However, these interaction methods that require the operation of certain devices are not convenient and smooth for the interaction, and also restrict the speed of the interaction. With the increasing progress and development of science and technology, the performance of computers is constantly improving, and its influence on people's lives is gradually expanding and deepening. High-performance computers urgently need more natural and smooth human-computer interaction methods, and at the same time provide the necessary conditions for realizing these human-computer interaction methods. The development of computer promotes the continuous innovation of human-computer interaction technology. Gesture rec...

AI Technical Summary

Problems solved by technology

In the process of algorithm execution, it generally needs to spend a lot of computer resources for calculation, which is not conducive to the real-time work of the system.

Images