This paper presents a solution for addressing this type of problem. In this study, we propose an ANN based salt and pepper filter (ANN-SPNF). Using the attributes of the nearest pixel values to the noisy pixel to be filtered, an ANN network model was constructed. In order to create the training set, the characteristics of the eight noiseless pixels closest to the noisy pixel were used. There are three attributes per pixel. These are the noise-free pixel value, the distance along the x-axis, and the distance along the y-axis. In the network input, 24 attributes are used in total. The network output is the original value of the pixel before the addition of noise. The ANN-SPNF was evaluated for all noise densities using medical images contaminated with salt and pepper noise. Using quality metrics, the outcomes of the algorithm technique were compared to the outcomes of other techniques. Compared to other approaches in the literature, the ANN-SPNF produced competitive results; it produced the finest results.

Image denoising, ANN, maching learning, Digital images. Salt and pepper noise, medical image