Quaternion based optical flow estimation for robust object tracking

We propose a quaternion optical flow algorithm for robust object tracking. Unlike previous works of color optical flow methods that treat color as separating channels, the proposed algorithm exploits quaternion representation of color and processes color as a holistic signal. In this way, it enables more accurate flow estimation at the pixel locations of spatial color variations, and reduces tracking errors by leaving more features points at their correct locations on the target. For successful and efficient object tracking, we also proposed a novel type of quaternion color corners that are reliable features during tracking. Together with grayscale corners, they form a good feature point set, especially when used with the proposed quaternion optical flow algorithm. We conduct a quantitative evaluation on publicly available dataset to verify the efficacy of the proposed algorithm. And object tracking experiments demonstrate that robust tracking can be achieved for real-time applications.