The velocity snake

We present a new active contour model for boundary tracking and motion estimation of non-rigid objects, which results from applying a velocity control to the class of elastodynamical contour models, known as snakes. The proposed control term is the outcome of an energy dissipation function which measures the difference between the contour velocity and the apparent velocity of the image (optical flow). Treating the image video-sequence as continuous measurements along time, it is shown that the proposed control results in an unbiased tracking, provided the initial contour is sufficiently close to the object boundary. This is in contrast to the original snake model which is proven to be biased due to the image (object) velocity; thus resulting in high sensitivity to clutter and numerical noise. The motion estimation further allows for position prediction of non-rigid boundaries. Based on the proposed control approach, we propose a new class of real time tracking contours, varying from models with batch-mode control estimation to models with real time adaptive controllers. The new tracking scheme was applied to boundary tracking of both rigid and non-rigid objects, resulting in unbiased tracking and robustness to image clutter and numerical noise.