Technique for mixed noise reduction based on support vector machine [image denoising]

Summary form only given. In this paper, we propose a new noise reduction method for images based on support vector machines (SVM). This method classifies pixels by their local features and processes them by a suitable method according to their features. In the proposed method, white Gaussian noise reduction with edge preservation is especially considered. The mixed noise reduction technique, based on a combination of the proposed method and an impulse noise reduction filter by using the SVM, is also described. Simulation results show the effectiveness of the proposed method for the reduction of white Gaussian noise and mixed noise with edge preservation.