A New Approach for Image Scaling Using DWT Interpolation

Image scaling is a technique which scale down or scale up the pictures or video frames to fit to the application. This work proposes a new scaling algorithm for the implementation of image scaling. The proposed scaling algorithm is an area pixel scaling algorithm consisting of a Discrete Wavelet Transform (DWT) based interpolation and bicubic interpolation. To achieve higher visual quality, DWT interpolation is carried out first to the gray scale values of image and then bicubic interpolation is performed. A simple Haar wavelet based DWT is implemented. DWT is based on sub band coding, which divides the image into four frequency quadrants. To reduce the artifacts, bicubic interpolation is performed to all the quadrants separately. This work can achieve an image quality by a factor more than 10 dB than the existing bilinear interpolation method. The average Peak Signal to Noise Ratio (PSNR) is more and the mean square error is less in this method. The image artifacts like blurring can be greatly reduced and checkerboard effects are eliminated in the proposed method, thus this approach is better than existing methods in visual quality. The simulation of the work is carried out in MATLAB R2013a. KeywordsImage scaling, Discrete wavelet Transform (DWT) interpolation, Bicubic interpolation, Peak Signal to Noise Ratio (PSNR).