Abstract: Diabetic is one the important cost related injuries. And also it gives the numerous eye problems for humans like diabetic retinopathy. In this research, a new method has been proposed to recognize the eye and also it has been analysed with other technique. Edge detection is one of the important modules of any image processing technique. In this research, researchers have been proposed the Canny Edge Detection technique based on Region of Interest (ROI) and Edge Length (EL) to recognize the human eye. And also Sobel Edge Detection technique is proposed to detect the edge of the diabetic retinopathy eye image. Both the technique have been compared and analysed. The performance of the both technique has been compared and verified with each other. After that it has been validated with existing standard values which was collected from the standard and famous eye hospital in coimbatore. In this research, Canny Edge Detection technique is a novel technique to identify the diabetic retinopathy and also the proposed technique shows significant results compared with Sobel Edge Detection technique.