Convolutional neural network and user habitual behavior analysis combination-based AR system gesture identification method

Abstract

The invention discloses a convolutional neural network and user habitual behavior analysis combination-based AR system gesture identification method. The method comprises the following steps of 1, collecting user habitual gesture images: randomly providing a group of gestures by a user, taking the group of the gestures as standard gestures, collecting images of the group of the gestures, and marking the images as a standard group; according to a gesture model graph of the standard group, establishing corresponding actual tag types; setting different tag types for triggering corresponding specified functions of an AR system; 2, detecting gesture region images: performing gesture region image detection on gesture images of the standard group, a training sample group and a test sample group to realize segmentation of skin color and non skin color regions in the images; and 3, realizing gesture feature identification by a convolutional neural network: designing a convolutional neural network preliminary structure model, training, testing and adjusting the convolutional neural network model by using sample data, and directly inputting a binary image to the convolutional neural network.The gesture identification accuracy is relatively high and the additional device cost is relatively low.

Description

technical field [0001] The invention relates to a gesture recognition method for an augmented reality (AR) system, in particular to a gesture recognition method based on a convolutional neural network combined with user habitual behavior analysis. Background technique [0002] In recent years, as artificial intelligence continues to enter people's field of vision, augmented reality technology (AR technology) has gradually become a hot topic. Augmented reality technology applies virtual information to the real world through computer technology, and the real environment and virtual objects are superimposed on the same screen or exist simultaneously in the same space in real time. Among them, human-computer interaction technology is particularly important. As a way of expression, gesture is usually regarded as one of the important means of human-computer interaction, and gesture recognition has also attracted many scholars' research. [0003] Chen Zhihua and others proposed a...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings of the existing AR system, such as low gesture recognition accuracy and high cost of additional equipment, the present invention proposes a convolutional neural network combined with user habitual behavior analysis with high gesture recognition accuracy and low additional equipment cost. AR system gesture recognition method

Images