Preprocessing and Segmentation of Retina Images for Blood Vessel Extraction

Abstract

In every field there is use of technology as part of in medical field lot of analysis is done using images. Processing of images will give good analysis when there is no noise or very less noise is present so processing retina images also gives correct results so we used different filters to find out which filter is suitable for pre-processing of retina images available in DRIVE database by computing mean square error (MSE) and Peak signal to noise ratio (PSNR) for different noises. After pre-processing images are segmented using discrete wavelet transform (DWT) and extracted blood vessel pixels are computed and compared with first observer result available in data base and results are very close to manual segmentation which is given DRIVE database.