In order to accurately obtain the birefringence dispersion of liquid crystal variable retarder (LCVR) with wavelength ranges from 414nm to 702nm. Here, we propose an improved fitting method of birefringence dispersion by using the theory of polarized light interference. The birefringence of LCVR in 478nm is 0.3863 based on the improved method. The experimental result is 0.3862 and the percentage error is 0.026%, compared to the 1.761% of the original fitting method. This work provides an improved way to character the birefringence dispersion of LCVR.
Iris recognition technology has advantages of high security and high stability and can realize non-contact and living recognition. For iris recognition, iris localization is the first step. The localization precision is assurance of high recognition rate. The existing iris localization method has shortage of slow speed. In this paper, an improved iris region localization method is proposed. Firstly, the images are preprocessed with Gaussian blur algorithm. Then the center of the pupil is coarsely located by Gaussian template convolution operation. Finally, the center and the outer edge of the iris are precisely located based on the Integral-differential operation. In experiments, CASIA iris database is used to test the improved iris region localization method. The results show that the location time can be less than 200 ms and the location precision can reach 98. 82%. This method’s usefulness is verified and can promote the application of the iris recognition