Image Quality Assessment Algorithms In Iris Recognition

At present, as an emerging of human biometrics for identification, iris recognition has become a research topic of Applied Mathematics, pattern recognition, information security and other cross-subject. In addition, iris image quality assessment is of the great significance for improving the iris recognition accuracy. So, With Matlab as the tool, main works about iris image quality assessment are described following:(1) The iris recognition principle and process were discussed, and the current domestic and foreign classical design methods of iris recognition system, which has four parts:image acquisition, image preprocessing, feature extraction, feature matching, were introduced. In addition, the actuality of iris image quality assessment algorithms was introduced briefly.(2) Based on the different factors affecting the image quality, a set of objective iris image quality assessment algorithm was proposed. Firstly, the distribution of histogram was used to judge whether the illumination is uniformity or not. Secondly, the morphological method was adopted to estimate the location of pupil. The specula reflection and the edges of iris were detected by gradient, gray and second-order derivatives of Gauss function. And, the linear interpolation and Direct Least Square were used to fill the specula reflection and fit ellipse pupil boundary, differently. Then, whether the pupil and iris is half-baked and anamorphic or not is known. Finally, the algorithms were designed in the frequency domain and the spatial domain separately to detect the eyelids, eyelashes, defocus and motion blurred images. Via analyzing the efficiency, a set of algorithms combining the frequency domain and the spatial domain was proposed.Carried on the batch experiments on database CASIA-IrisV3-Lamp, the proposed algorithms were proved to effectively distinguish the unqualified images from the ideal images, and the ones failed to pass the assessment system always can not be used for iris recognition. That is to say, iris image quality assessment greatly enhanced the practicability of iris recognition system and avoided noneffective account, then validated that iris image quality assessment is of the great significance for iris recognition system.