Pseudocepstrum for assessing stereo quality of retinal images

Diabetic retinopathy (DR) is the primary cause of visual loss in the west, although 95% of visual loss cases due to DR can be prevented if detected and treated early. Early treatment of diabetic retinopathy study (ETDRS) has been set by the National Institute of Health (NIH) to assign a severity level based on evaluation of stereo retinal images of people suffering from DR, and was described as the Gold standard for early detection and treatment. To produce accurate results, the implementation of the standard depends on high resolution color stereoscopic images of the retina that meets certain quality criteria such as clarity & focus, field definition, and stereo effect. A fast and robust image quality assessment (IQA) tool can is imperative to efficient and accurate disease diagnosis based on ETDRS. Specifically, this paper focuses on the stereo quality of the image pair due to its importance in the grading process, and we propose the use of the cepstrum or pseudocepstrum filtering to develop an efficient IQA tool for assessing the stereo quality of the retinal image pair.