AN AUTOMATED TECHNIQUE FOR DIABETIC DAMAGE DETECTION THROUGH BLOOD VESSEL SEGMENTATION IN RETINAL IMAGES

摘要

This paper presents a new supervised method for blood vessel detection in digital retinal images. This method uses a neural network scheme for pixel classification and computes a 7-D vector composed of gray-level and moment invariants based features for pixel representation. The method was evaluated on the publicly available DRIVE and STARE databases, widely used for this purpose, since they contain retinal images where the vascular structure, vessel endothelium and fund us features has been precisely marked by experts. Analysis of the designed technique on different kind of images that is on DRVIE database and compare to existing blood vessel segmentation techniques. Method performance on both sets of test images is better than other existing solutions in literature. The method proves accurate for blood vessel detection and its performance analyzed segmentation approaches. With this simplicity and fast implementation, make this blood vessel segmentation proposal suitable for retinal image computer analyses such as automated screening for early diabetic retinopathy detection.