Restoration of Texture for Gray Scale Images by Error Reduction Algorithm using FFT Estimation
Harish Velingkar, Veena Devi · 2014
A method for reconstruction of missing texture using an iterative technique of Error Reduction(ER) algorithm & fast Fourier transform estimation has been explained in this brief. In our method we try to recover the missing texture patches in the image with the help of the errors converged in the ER algorithm using the known intensities. We initially identify two regions, one with known texture characteristics and the other with texture degradation ie. with unknown texture characteristics. We then break down both these regions into a number of smaller blocks and try to estimate the unknown texture characteristics by iterative comparison of the FFT magnitude with each of the known blocks. We further recover the phase of the original image using ER algorithm again and finally implement a circular filter to obtain the required results. Metrics such as SSIM,MSE & PSNR will be used to determine the amount of reconstruction obtained using the above method.