Image spatial resolution enhancement: A novel wavelet approach

Enhancement of image spatial resolution is a widespread topic in the field of image processing and there exists various solutions with diverse approaches. But in the last few years wavelet based approaches show good results in image resolution enhancement. In this paper, we present another new wavelet based image resolution enhancement technique where discrete wavelet transformation (DWT) is used to decompose the low resolution (LR) image into frequency subbands (horizontal, vertical and diagonal). All these frequency subbands are binarized and then interpolated using lanczos interpolation. At the same time Laplacian filter, Sobel operators and lanczos interpolation are used to estimate another set of binarized horizontal and vertical subbands. Finally these pairs of binarized horizontal and vertical subbands are added (bitwise OR) and used into the inverse DWT process for generating a new high resolution (HR) image. This new technique has been tested on four famous test images Lena, Baboon, Elaine and Peppers. Experimental results show that the proposed technique outperforms conventional and state-of-the-art techniques in terms of Peak Signal to Noise Ratio (PSNR), Root Mean Square Error (RMSE), entropy and Universal Quality Index (UQI) value. In addition behavior of the proposed technique for different wavelets are simulated and discussed.