Image Quality Improvement by Multi-Scale Dictionary Method

This paper presents the comparison of different techniques of Super-Resolution (SR). Multi dictionary technique is compared with SR in the frequency domain and Interpolation-restoration.This paper shows how multiscale dictionary gives high quality SR images than the other two techniques. Frequency domain technique fails to presence of outliers due to motion errors, inaccurate blur models, noise, moving objects, motion blur etc. Interpolation restoration fails to do intensive computation due to large number of unknowns. Multiscale dictionary overcome this problem by integrating local and non-local priors, and gives better outliers and intensive computation.