Performance Analysis of Image Reconstruction Based on Modified Radon Transform with Wavelet Transform

Abstract

In many medical imaging modalities, image reconstruction is a key component. Radon Transform is mostly used for reconstruction due to its ability to map from the image space to line parameter space. The goal of this work is to perform another alternative to remove blurring on reconstructed images which is obtained by the Classic Radon Transform. Wavelet Transform, which provides time and frequency representation, allows complex information such as images to be separated into elementary forms at different positions and scales and reconstructed with good precision. In consequence, we have an additional chance to distinguish image content and noise. By combining benefits both Radon Transform and Wavelet Transform, it becomes one of many ways to reconstruct an image. In addition, filters are applied on reconstructed images to improve their performance. The quality of the images is evaluated by Peak Signal Noise to Ratio (PSNR) and Structural Similarity (SSIM) to find the performance of the reconstructed images. Numerical results show that the Modified Radon Transform with Wavelet Transform plus wiener filter is more effective than the Classic Radon Transform.