An Approach for Validation of Improved FeaturePoint Detection in Human Retinal Recognition

Conference: Third International Conference on Current Trends in Engineering Science and Technology
Author(s): Prashantha G R, Chandrashekar M Patil Year: 2017
Grenze ID: 02.ICCTEST.2017.1.179 Page: 1038-1045

Abstract

The retinal recognition is yet another step in biometric, as a very rare but the most accurate authentication technique. In this paper a windowing technique for feature extraction and feature point detection in order to improve the accuracy of retinal recognition. The proposed paper contains distinct property exceptionally low FAR and FRR. This paper presents a four stages feature extraction and feature point detection using retinal vascular pattern of Human retina. In first and second stage, Image acquisition, is the process of obtaining an image from the source. Image preprocessing is the image resizing that is before resizing the image was shown at 67% instead of the normal 100%. It involves four process, RGB to gray scale and green channel image selection, Histogram equalization, Noise removal, filtering, sharpening and smoothing in third step. In forth stage, it performs feature extraction and feature detection using Morphological structuring and FAST (Features from Accelerated Segment Test) algorithm. The proposed method is tested on available databases DRIVE. Experimental analysis of given approach shows a significant decrease of false feature point.

<< BACK

ICCTEST - 2017