Context Aware CNN approach to denoise Salt and Pepper Images

Salt and pepper noise is a type of impulse noise that affects digital images. It appears as bright or dark pixels that are randomly distributed throughout the image. Salt and pepper noise can be caused by a variety of factors, such as sensor defects, transmission errors, and software errors. Denoising salt and pepper noise is important because it can significantly improve the quality of an image, making it easier to interpret and process. It can also improve the performance of image processing tasks such as edge detection, segmentation, and classification. There are a number of methods for denoising salt and pepper noise, including median filtering, adaptive median filtering, morphological filtering, and deep learning-based methods. The following paper discusses some of the latest methods which are used to denoise SAP affected images and identifies some of the shortcomings of the current methods. This forms the basis of the research which has been subsequently carried out to design a new method of denoising SAP affected images using deep learning. This paper proposes a new model to denoise SAP affected images using custom layers from the Tensorflow library along with Convolutional Neural Networks, specifically the DnCNN model.