A new method of image denoising based on cellular neural networks

摘要

This paper presents an edge constraint adaptive filtering algorithm based on cellular neural networks for imagedenoising. In the process of designing the three templates separately in cellular neural networks, the control template referencesthe advantage of spatial filtering denoising. It resembles a spatial domain denoising filter. The feedback template sets as a matrixwhich generated by a high pass filter to achieve edge preservation. The proposed method can not only achieve denoising, but alsoprotect edges in an image. In the process of designing the threshold template, we use the different gray levels in an image toachieve the threshold adjustment adaptively. The experiment simulation results show that this algorithm is effective. Its denoisingeffect is much better than the mean filtering, median filtering, Gaussian filtering and the non local means method. And comparedwith the anisotropic diffusion algorithm, this algorithm is also better for the impulsive noise (salt & pepper noise), the Poissonnoise and the comprehensive noise denoising. Due to the parallelism and possible hardware implementation of cellular neuralnetwork, it can achieve real time image denoising, which has a good application prospect.