Real-time user interface using particle filter with integral histogram

To apply a vision technique to real-time user interface, we suggest a fast and effective method of tracking human limbs and recognizing poses, with the kinematics chain model. The proposed method is able to evaluate the observation likelihood between particle samples and reference histograms in high speed with the integral histogram. Results demonstrate that using the integral histogram is faster than the existing color histogram method as evidenced by the increase in the number and size of samples. In addition, robust results are realized during evaluation involving overlapping objects. The proposed user interface system is ideally suited for gaming devices that utilize user gestures and poses as control input in real-time. Experiments for the real-time game demonstrate the robustness of the proposed method.