KHGI: Kinect-based Human Gait identification using Statistical Moments

As one of the biometric recognition systems, gait recognition aims to recognize the human gait style on-the-go when they walk. The Kinect sensor is a new camera originally designed for game play, which recently has been focused on by researchers, especially for gait based recognition methods. This study focuses on using Kinect sensor for extracting gait signatures to be used for gait recognition purposes. Moreover a new feature set is extracted by calculating Euclidean distance between each pair of 20 joint points, followed by the application of various statistical moments during one gait cycle, called the Euclidean Distance Feature (EDF). The extracted feature vector is a high dimensional feature set. Therefore, we need to decrease the dimensions of the feature set, hence we applied, Principal Component Analysis followed by Linear Discriminant Analysis. Finally, the Linear Discriminant Classifier (LDC) and Nearest Neighbors (NN) are used separately as classification methods. The results indicated that the proposed method achieves a significant result and provides a 92% recognition accuracy.