Research On Personal Identification System Based On Function Of Kinect Skeletal Tracking

Whether a personal identification system can quickly and effectively used for identification becomes increasingly important in the current society.But because of the physical likeness and professional salon technology,face recognition may could not be used to recognize well the characters.Personal identification system based on function of kinect skeletal tracking not only filled up the technology vacancies but also provided more basis support in the study of reorganization.Different people have different arm length,led length,even twins can't be the same,which can be used as a for personal identification.What is more?Personal identification system based on function of kinect skeletal tracking has many advantages compared to other recognition's methods,including having strong individual difference and stability,having intuition and convenience.Considering the fact,the paper presented an individual identification system based on skeletal features.In the process,back-propagation artificial neural networks and support vector machines was introduced to improve the recognition's quality.We designed a windows form application interface which aimed to provide users with more convenient intuitive ways to use devices.Users can click through to the corresponding button,into a different processing interface.As time limited to study inevitable shortcomings,we conducted the study to a certain reflection with a view to further explore in future.The Kinect for Xbox 360 with three different types of cameras(motion sensing device RGB camera,infrared camera and Infrared ray CMOS camera)is a key measurement equipment in this testing.Here shows the experiment primary steps of the recognition system.Establish system develop environment,complete system frame and program.Firstly,In the VS2010 environment through various programming method can realize Kinect,driver to body's 20 joints acquisition and playback from a C++ programming perspective.It loads depth sensor in the way of "Light Coding" within a Kinect,so that users can get the depth data and projection of the infrared pattern,and obtain depth information.So the images with 3D depth will be obtained.This was the first step to get skeletal features.Then,the target person was separated out only by image segmentation module.Analysed the pixel of the photos and figured out all parts of the body.Finally recording and calculating the body's skeletal data were necessary.We chosen the valuable data as the characteristic data and used BP-ANN classifier to train it.In this way,personal identification system based on function of kinect skeletal tracking could be finished.