Human Recognition based on Retinal Bifurcations and Modified Correlation Function
This addresses security needs for controlled access in secure places, but appears incremental as it builds on existing retinal recognition methods.
The paper tackles human recognition for high-security access control by proposing a novel identification method using retinal images, achieving 99.34% accuracy on a dataset of 200 images.
Nowadays high security is an important issue for most of the secure places and recent advances increase the needs of high-security systems. Therefore, needs to high security for controlling and permitting the allowable people to enter the high secure places, increases and extends the use of conventional recognition methods. Therefore, a novel identification method using retinal images is proposed in this paper. For this purpose, new mathematical functions are applied on corners and bifurcations. To evaluate the proposed method we use 40 retinal images from the DRIVE database, 20 normal retinal image from STARE database and 140 normal retinal images from local collected database and the accuracy rate is 99.34 percent.