Abstract: Aiming at the problems existing in the process of iris quality evaluation, such as how to select appropriate evaluation factors, how to reduce the calculation amount of evaluation factors, and how to effectively integrate evaluation factors, we proposed an iris image quality evaluation method based on genetic algorithm-support vector machine (GA-SVM) model and multi-measure evaluation indexes. Firstly, the definition quality of iris images was evaluated, and fuzzy images were roughly screened out. Secondly, four evaluation indexes were selected and GA-SVM model was used to effectively fuse the evaluation index values to comprehensively evaluate the quality of iris images. Finally, the method was verified in JLU-6.0 iris library and compared with other classical evaluation methods. Experimental results show that this method can improve the survival rate of available iris images, achieve better recognition accuracy, and improve the running speed of the system.

Key words: iris image quality evaluation, support vector machine (SVM), genetic algorithm, multi-index fusion, binary classification