Impulse Noise Removal Using Adaptive Radial Basis Function Interpolation

A novel adaptive radial basis function interpolation-based impulse noise removal algorithm is introduced in this manuscript. This approach consists of two stages: noisy pixel detection and correction. In former step, the noise-affected pixels in an image are detected, and in the latter step, the noisy pixels are restored by adaptive radial basis function-based interpolation scheme. The radial basis function interpolation scheme is used to estimate the unknown noisy pixel value from the noise-free known neighboring pixel values. For both noisy pixel detection and correction, a center sliding window is considered at each pixel location. The proposed approach is experimented on some benchmark data sets, and its performance is evaluated using five performance evaluation measures: PSNR, MSSIM, IEF, correlation factor, and NSER on different test images by comparing it against sixteen different state-of-the-art techniques. It is found that the proposed approach gives better results than the sixteen different state-of-the-art techniques.