Performance comparisons of interpolation algorithms for 2D images
C.X. Wang, W.E. Snyder, Griff L. Bilbro · 2002
The current study concentrates on performance comparison of the MAP based algorithm with several other well known image interpolation methods such pixel replication, linear interpolation, and spectrum extrapolation. Analyses including mean square error, histogram analysis, and the matched filtering analysis were carried out. From the experimental results, the MAP based algorithm produced zoomed images with a superior quality, in comparison with existing algorithms. The image produced by the MAP based algorithm have a lower MSE and its histogram suffers less distortion. For the matched filtering analysis, the MAP algorithm produces a high template correlation for the target object.