Color image interpolation in the DCT domain using a wavelet-based differential value

Abstract

We propose a new image-interpolation technique using a combination of an adaptive directional wavelet transform (ADWT) and discrete cosine transform (DCT). In the proposed method, we use ADWT to decompose the low-resolution image into different frequency subband images. The high-frequency subband images are further transformed to DCT coefficients, and the zero-padding method is used for interpolation. Simultaneously, the low-frequency subband image is replaced by the original low-resolution image. Finally, we generate the interpolated image by combining the original low-resoultion image and the interpolated subband images by using inverse DWT. Experimental results demonstrate that the proposed algorithm yields better quality images in terms of subjective and objective quality metrics compared to the other methods considered in this paper.