Measurement of ocular torsion using iterative Lucas-Kanade optical flow method

This paper presents a new method for measuring ocular torsion using the optical flow. Iris image is cropped and transformed into rectangular image to make a orientation invariant image. Feature points were selected at iris region from a reference and a target image, and then shift of each feature was calculated using iterative Lucas-Kanade method. The feature points were selected according to the strength of corner on the iris image. The accuracy of the algorithm was tested using printed eye image and compared with traditional cross-correlation method. Measurement error was less than 0.15 degree.