The Algorithm of Iris Image Quality Evaluation

In this paper, we judge whether an iris image has quality problem from two aspects. On one hand, we analyze the higher spatial frequency sub-band image of the original iris normalization image and determine whether this image has eyelid and lash occlusion problem according to the number of high value pixels. On the other hand, we determine whether an iris image is clear enough though computing the variance of the wavelet coefficients. Through our experiments, this algorithm is proved to be universal and efficient.