A Facial Features Matching Method with Age Robustness Based on SIFT Algorithm

Facial change with age is always a problem in facial features matching for photos of ID cards. In order to solve this problem, a novel facial features matching method based on SIFT algorithm is proposed. Inspired by attention mechanism of the visual system, the facial features focuses on the four sub regions (left eye, right eye, nose and mouth), ignoring other unimportant regions, and different weights are used to each region according to the contribution of each region to matching accuracy. The experimental results show that the proposed method can effectively reduce the influence that age change lead to a sharp decline in matching accuracy, and this algorithm can significantly reduce the time of features matching.