Image Interpolation Using Spatial And Spectral Correlations

Abstract

This paper presents a new method to interpolate digital color images by exploiting the spatial and spectral correlations inherent in the images. The proposed method involves two main steps: interpolating green image plane solely based on the local spatial correlation within the color plane, and interpolating red and blue image planes based on the spectral correlation between different color image planes. The performance of the proposed algorithm is evaluated by interpolating down-sampled color images back to their original image sizes and comparing the results with the original images as well as those generated by other existing image interpolation schemes. Experimental results show that the proposed method can better preserve valuable image attributes: edge sharpness and color consistency, which have great influence on the perceived visual quality of the interpolated images.