Enhancing the Still Image Using Super Resolution Techniques: A Review

Devidas D. Dighe; Gajanan K. Kharate; Varsha H. Patil

doi:10.37591/jomtra.v2i2.425

Abstract

Super Resolution (SR) refers to the reconstruction of images that are visually superior to the original low resolution (LR) images by bandwidth extrapolation beyond the pass band of the imaging system. Tsai and Hunag were the first to consider the problem of SR in 1984. Onwards over three decade various researchers contributed in the field of SR but all are intuitive SR mechanisms. This paper reviews the recent SR techniques. From the observations, the SR techniques are classified as; frequency domain or spatial domain techniques, but also need to classify SR techniques based on SR using multiple LR or single LR image(s). Survey carried by us revels that, the researches on SR reconstruction mainly considered the linear degraded model, results provided are mostly based on subjective measurements, and it is difficult to find an unbiased comparison. There must be considerations for number of available LR or HR image(s) for selection of appropriate SR technique. Hence, there is need to provide a clear method of comparing different implementations suitability, so one has to implement SR method based on problem model which can be generalized to all SR reconstruction problems.Keywords: Super Resolution, Registration, Reconstruction, Subpixel, Aliasing, BlurringCite this ArticleDevidas D. Dighe, Gajanan K. Kharate, Varsha H. Patil, Enhancing the Still Image Using Super Resolution Techniques: A Review. Journal of Multimedia Technology & Recent Advancements. 2015; 2(2): 35–51p.