Two dimensional autoregressive modeling-based interpolation algorithms for image super-resolution: A comparison study

Image interpolation is a key technique of image super-resolution. Four two dimensional (2-D) autoregressive (AR) modeling-based image interpolation algorithms have been reported to have better performance in edge and texture preservation than conventional image polynomial interpolation algorithms. However, there is lack of performance comparison among them. For super-resolution reconstruction quality, this paper is going to fill up the gap by a comparison study on the four 2-D AR modeling-based interpolation methods: novel edge-directed interpolation (NEDI), soft-decision adaptive interpolation (SAI), sparse representation interpolation with nonlocal autoregressive modeling (SR-NARM), and adaptive super-pixel-guided AR modeling (ASARM). Furthermore, the four interpolation algorithms are compared in the light of peak signal to noise ratio, feature similarity index, mean squared error and structural similarity index. From comparative results we observe that ASARM method has relatively better performance than other three methods but is more time-consuming than the NEDI and SAI methods.