PROJECT TITLE :

Hand Posture Recognition Using a Three-Dimensional Light Field Camera

ABSTRACT:

This paper used a light-weight field camera to capture 2 types of 3D hand posture depth images, contour hand posture and solid hand posture, while not a sophisticated setup and superfluous preprocess. The images were captured under 2 recognition conditions: in-plane and out-of-plane rotations. The posture recognition features 2 methods: the primary methodology is the principal component analysis (PCA), which is used to obtain the specified feature vectors, related to the k-nearest neighbor (k-NN) algorithm as a classifier; the second method involves using 2D optimal-PCA (2DOPCA) combined with a genetic algorithm (GA) for feature selection, and therefore the Mahalanobis distance is then used for classification. The variations within the test images embody in-plane rotation, out-of-plane rotation, and Gaussian noise added to simulate the lighting interference during a real state of affairs. The results showed that the PCA combined with the k-NN yields high recognition rates for grayscale contour images with in-plane and out-of-plane rotations and color solid posture images with in-plane rotation. For color solid posture images with out-of-plane rotation, the projection color house was combined with the PCA and k-NN ways to obtain high recognition rates. Moreover, the contour and color solid posture pictures with noise additional than 5% require the 2DOPCA combined with the GA to get a satisfactory result and maintain recognition rate stability.