Human Identification Based on Gait

Gait based human recognition system is most important and attractive method of biometrics. Gait means the way of walking. Person video is captured in a particular view (90 degree in our case) which is side view from distance and provides more efficient means of biometric verification here we use gait as biometric type. In this dissertation, we propose an efficient algorithm which works on angle based technique and human recognition is done and implementation of this work is done. In our work initially video is captured and then converted into frames after that feature extraction is done from given images for one cycle. Here in this work three lower body parts are taken for recognition and a correlation of triangle is derived. Using cosine and sine formula each inner angle of triangle is calculated and mean value of each angle of triangle is stored in database for identification. If person is match with database than it gives result that person is found and identification is done. The gait algorithm implemented in MATLAB. To accomplish this research work we used CASIA ¿A¿ set (which gives side view images) for our result.