Skeleton Recognition Technology Research Based On Gait

Gait recognition technology is a biometric technology, we can get acquisition sequenceof images of the human body during walking by using the fixed camera, and then processedthrough the relevant algorithm compared with the stored gait data, this method used byidentity identification. Gait recognition as a biometric technology, its unique advantages thatother biometrics do not have, so it have a potential to identify the remote video image orlow-quality video, and the gait is not easy to hide or disguise. With a full range of smartstations, banks, airports and other sensitive occasions require a wide range of enhance themonitoring system, remote identity authentication system to the computer vision researchershave generated great interest. The characteristic of the gait we can get by long-rangeenvironmental, so from the point of view of intelligent video surveillance, gait recognitionvery broad application prospects. Gait recognition has become a new field of study, by peoplewalking gait characteristics of the process for identification. Watch this topic, this paperlaunched a study of the following aspects:gait sequence extraction and building database, this article get the gait image form aviformat video, and save it as a JPG picture format, in order to in-depth analysis of images, andcollect six young people in indoor gait video sequence in indoor and set up a database.extraction of gait contour map, including two parts, they are background modeling andmotion segmentation. Set up a modeling by GMM, by modeling can get the video backgroundimage, background subtraction method to obtain the corresponding motion detection site, andthen use of morphological detection and edge detection, image binaryzation to obtain a goodgait image. From the experiment, the contours of the human body get clearer. Using templatemethod. This method application is essential to skeleton model gait contour map, and extractcharacteristic information from human gait. The first to use the skeleton algorithm based ondistance transform to extract the human skeleton; extract by Corrosion and restore technology torso and legs of the human body as a whole skeleton and skeleton branches; Finally, the maintrunk and lower limb skeleton cyclical swing characteristics as gait recognition feature tohuman gait identify.Experimental results show that this method has a high accuracy rate in fixed indoor lighting conditions.And in the case of different perspectives, the skeleton information Perspective correction can also be acertain extent to improve the rate of correct identification. However, in practical applications still have toconsider many factors, such as environmental adaptability, three-dimensional perspective of change andother issues.