A New image Interpolation using the Bilateral Filter

Abstract

Several factors interfere with the quality of image such as aliasing, noise, artifact and blurring. These factors can cause the degradation of the image especially in the edge regions. In order to reduce their effect, it is necessary to choose a robust interpolation method for preserving the edges and textures, and improving the quality of the output image. In this paper, a novel method of image reconstruction is proposed, based on image interpolation using the bilateral filter, which consists in estimating the edge-directions by the bilateral filter, then interpolating the missing pixels along the detected edge direction using a cubic convolution. Firstly, We generate the low-resolution image, by performing directly a
down-sampling by a factor of two to the original gray high-resolution image , and then we apply our proposed interpolation method to generate the HR image. Finally, this method is implemented in MATLAB and tested on a large number of gray images, and afterwards an accurate comparison is made with the other conventional interpolation methods, both objectively and subjectively. Statistical simulation results demonstrate that the proposed
method outperforms the conventional interpolation methods in terms of quantitative evaluation PSNR, and two perceptual quality metrics SSIM, FSIM, and produces sharper edges and cleaner image details in terms of subjective visual quality of the reconstructed images.