Blind deconvolution and retinal abnormality detection in blurred retinal images

Abstract

In this paper, a new technique is presented to enhance the blurred images obtained from retinal imaging; both regular as well as Fluorescein Angiography. While in most of the cases, the image produced is quite clean and can be easily used by the ophthalmologists, there are many cases in which these images come out to be very blurred due to the disease in the eye such a cataract etc...In such cases, having an enhanced image can enable the doctors to start the appropriate treatment for the underlying disease. The proposed technique utilizes the Blind Deconvolution approach using Maximum Likelihood Estimation technique. Further post-processing steps have been proposed as well to extract/classify specific regions from the image automatically to assist the doctors in visualizing these regions related to very specific diseases, for instance, blood vessels and hemorrhages for diabetes, etc. The post-processing steps include Image color space conversions, thresholding, Region Growing, and Edge detection.