Supervised Learning Method for Image Restoration of Natural Images

D. Khalandar Basha and Dr.T. Venkateswarlu

The restoration of image is one of the thirst area for doing research. The goal of image restoration achieved by more amount of noise removal from the degraded image. For image restoration various algorithms proposed. In this paper image restoration is done by using support vector machine (SVM) algorithm for regression. In the proposed algorithm Support vector machines is used for regression. The kernel used is radial basis function (RBF) type. The proposed algorithm is applied for various images. The restoration is done for the images corrupted by Gaussian noise and salt – pepper noise. The amount of noise is varied from 10% to 50%. The proposed algorithm is able to restore image effectively for both noises. The performance of the algorithm is measured by MSE, RMSE, PSNR, SSIM and FSIM.

Volume 11 | Issue 7

Pages: 151-157