A Survey on Adaptive Image Interpolation based on Quantitative Measures

Abstract

Image Interpolation is a method of constructing data points within range of a discrete set of known data points. It is used to enlarge and compress the images. In non-adaptive interpolation, same operation is carried out throughout the pixels whereas in adaptive interpolation different operation is carried out throughout the pixels. Therefore, the adaptive interpolation technique improves the quality of the image. In this paper, different types of adaptive interpolation algorithm have been analyzed and compared. These adaptive interpolation algorithms are analyzed based on their quantitative measure such as PSNR (peak signal to noise ratio) and MSSIM (mean structural similarity index measure). From the survey followed by analysis, it is observed that the non-linear cell average interpolation scheme reduces artifacts and increases the smoothness at edges to improve the quantitative measure of interpolated image. This local weighted interpolation algorithm has a higher PSNR value of about 44.79 dB. Here is an improvement of PSNR value from 1.94 dB to 26.29 dB when compared with other interpolation methods. This survey can suggest that the quantitative measure of image interpolation can be improved by reducing jagging and blurring effect of the images using appropriate filters.