Research The Technology Of Hand Gesture Feature Extraction And Recognition

With the progress of the times and technology, more and more people are close friends to human-computer interaction. Among them, the gesture to the expression, with its advantages of simple easy to operate becomes an important technology in the field of human-computer interaction, The gesture becomes a key technology of human-computer interaction.Gesture recognition is one of the hot topics of current research, it involves multiple disciplines of image processing, pattern recognition, computer vision and many other disciplines. It is gradually applied to entertainment, to experience more convenient Static gesture is the focus of this article, The gesture recognition include gesture pretreatment, gesture detection segmentation, gesture detection segmentation, feature extraction and gesture recognition. Among all, how to extract effective gesture feature is a key part of gesture recognition.Firstly, the gesture image preprocessing and segmentation. Preprocessing part is normalized, including the scale and intensity normalized, so that the image has a "uniform standards" to reduce the differences due to excessive light or affect the size difference of recognition rate. Gesture segmentation part used is based on YCbCr color space and binary morphology Gaussian model combining algorithm compared with other methods to achieve good results.Secondly, the gesture feature extraction has done exploratory study, the Multi-Block Local Binary Pattern texture feature extraction applied to the gesture, which is improved on Local Binary Pattern basis.This improved effective use of local relevance and ability to describe the details of the strong advantages of information between pixels, enhanced resistance to noise algorithm, and can be divided according to the sub-blocks of different sizes, get a different scale gestures feature, realize the gesture information more complete description, and the standard database extraction experiments were carried out to verify the effectiveness of the texture feature.Finally, use Support Vector Machine method for classifying different gestures training, and use Massey University gesture database for standard database to compare experiments, it shows that this method achieved higher recognition rate and achieve a better effect.