FAST EYE DETECTION METHOD USING BLOCK CONTRAST AND SYMMETRY IN MOBILE DEVICE

摘要

Disclosed is a method for detecting the positions of the eyes at a high speed using block contrast and symmetry between a pupil block and surrounding blocks in a mobile device. The method comprises the following steps of: detecting a face area from an input image, by using Haar-like features, an AdaBoost algorithm and an adaptive template matching; determining an eye detecting area for right and left eyes using the detected face area; extracting candidate areas for eyes using a feature that a pupil block is relatively dark and symmetric is good compared to other surrounding blocks in an integral image in each of the eye-detecting area; selecting center pixels having block contrasts of the three highest ranks for each left and right side from the eyes candidate areas as pupil candidate points; inspecting that a distance between the pupil candidate points which are capable of being left-right paired is within a predetermined range of an upper limit distance to a lower limit distance and an absolute value of a gradient between two pupils is an angle equal to or less than a predetermined value at the same time, and determining the paired pupil candidate points which meets the criterion as the positions of the eyes.