A REAL TIME HAND TRACKING METHOD USING MOVING EDGE ORIENTED MEAN SHIFT ALGORITHM
Jaeho Lee, Jaekwang Kim, Changick Kim · International Conference on Multimedia and Information Technology · 2011
The color-histogram based mean shift algorithm has been used in various object tracking tasks. However, it often shows many false-positives when the background color is similar to that of the target object. In this paper, we present a robust method to track the hand region in a static camera scene. Our goal is to track the hand region irrespective of hand shapes or background color. To prevent errors due to the similar color distribution, we combine moving edge feature with the background-weighted color histogram. The weight terms of two features are adaptively selected at each frame according to the color similarity between the hand and the background region. Experiments on various test sequences which include a variety of hand shapes and colors prove the efficiency and robustness of the proposed method.