The Research Of Gait Recognition Based On Hidden Markov Model

This thesis researchs on gait recognition based on Hidden Markov Models (HMM). Human gait recognition is the process of identifying individuals by their walking manners and is attractive in pattern recognition and image processing. It can be applied to security system, human ID management, digital surveillance and so on.Generally, gait recognition consists of three parts: preprocessing of gait sequences, feature extraction and classification. Feature extraction is the most important thing, on which our study focuses.The preprocessing of gait sequences is extract human motion from gait video. It mainly includes background model, foreground detection and morphological postprocessing. The segmentations of gait silhouettes have very important influence on feature extraction and subject classification.In the feature extraction step, this thesis divides the silhouette into seven parts according to the human proportion, fits each part with an ellipse and extracts the parameters of the ellipse as the silhouette feature. It uses Kmeans algorithm to obtain 5 key frames of a gait cycle, and calculates the distances between every frame and the key frames to further reduce the observation vector dimension. These gait feature sequences can train and obtain a continuous HMM for every person, therefore the 5 key frames and the obtained HMM can represent each person's gait sequence. Finally, in the classification step, we calculate the output probability of the unknown gait sequence to all HMM and choose the person corresponding to maximum output probability as its classification identification.This method captures human gait structure and the transitional information, the HMM statistical property decides that it has very good robust in the gait description and recognition. At last, our experiments show that HMM is an effective method in gait recognition field and has a broad application prospect.