Method for determining quality of iris image based on machine learning

Abstract

The invention discloses a method for determining the quality of an iris image based on machine learning. The method comprises the following steps of: pre-processing the iris image; extracting quality factors of the iris image; fitting probability density functions of positive and negative samples of single quality factors by using multiple Gaussian models; performing fusion by using an improved Neyman-Pearson method to acquire the quality score of the iris image; and determining an optimal quality level by a hypothesis testing method. The invention provides a robust detecting method aiming at defocusing, motion blur and heterotropia; and multiple quality factors are fused by introducing the Neyman-Pearson method to form the quality score, and the image quality level with statistical significance is acquired by the hypothesis testing method. The method can be used for quality determination when the iris image is acquired and performance prediction aiming at a recognition algorithm.

Description

technical field [0001] The invention relates to computer vision, digital image processing and pattern recognition, in particular to a method for determining the quality of iris images based on machine learning. Background technique [0002] As a reliable identification technology, iris recognition has been widely used, such as identification in airports, customs, and financial institutions. However, due to the limitation of the imaging range of the iris sensor, it is difficult to obtain iris images that meet the requirements without the cooperation of the user, and the image quality has become a bottleneck in the development of iris recognition. In the process of image acquisition, a large number of iris images of different quality are added to the recognition sequence. It is precisely because of the addition of low-quality iris images that the performance of the iris recognition system is greatly reduced. With the development of large-scale and long-distance iris recogniti...

AI Technical Summary

Problems solved by technology

However, they only considered the effect of a single factor, and ignored the influence of multi-factor joint action on iris image quality. Relatively speaking, there are not many multi-factor joint quality determination methods, and the premise of multi-factor joint is to correctly evaluate each iris image quality. In addition, the quality score after fusion is too subdivided, resulting in no distinction between quality levels. For details, see the literature [3] N.D.Kalka, et al., "Estimating and Fusing Quality Factors for Iris Biometric Images , "IEEE Trans. on Systems, Man and Cybernetics, Part A: Systems and Humans, vol.40, pp.509-524, 2010

[0004] To sum up, the existing iris quality determination algorithm cannot accurately reflect the overall quality of the iris image, and how to obtain a quality score that correctly reflects the quality of the iris image is still a hot issue

Images