Fast motion estimation using bi-directional gradient methods
Dept. of Electrical Engineering Systems, Tel-Aviv University
Gradient based motion estimation techniques are considered to be in the heart of state-of-the-art registration algorithms, being able to account for both pixel and subpixel registration and to handle various motion models (translation, rotation, affine, projective). These methods estimate the motion between two images based on the local changes in the image intensities while assuming image smoothness. This work offers two main contributions: (i) Enhancement of the Gradient methods technique by introducing two new bi-directional formulations of the Gradient methods, which improve the convergence properties for large motions. (ii) We present an analytical convergence analysis of the GM and its properties. Experimental results demonstrate the applicability of these algorithms to real images.
Yosi. Keller is a PhD student at the Dept. of Electrical Engineering Systems, Tel-Aviv University majoring in low-level computer vision and image reconstruction algorithms.