Enhanced example-based image super-resolution for remote diagnosis of diabetic retinopathy in Thailand

This paper proposes a new image super-resolution algorithm to solve the problem of unclear retinal images sent from remote diagnosis of diabetic retinopathy in Thailand. The proposed algorithm is developed from Example-Based Learning algorithm. The functioning ability of the proposed algorithm, however, was enhanced to create a sharper image of which structure are well preserved. The systematic combination of the two algorithms brought about a new algorithm that can be effectively used with the retinal image. Experimental results showed that, the new algorithm has better performance than the existing algorithm namely Bilinear, NEDI and NARM.