Abstract:

A new denoising algorithm based on pulse coupled neural network (PCNN) system and local Wiener filter is presented. The mechanism of synchronous pulse of PCNN is employed to revise the image wavelet coefficients, according to the feather of different image pixels. Also, an adaptive linking coefficient β is proposed. Based on the improved wavelet coefficients, the local Wiener filter is used to estimate signal variance better. Experiments show the proposed scheme can get higher peak signaltonoise ratio than NeighShrink and Wiener algorithm, and better vision effect. 

Key words: pulse coupled neural network, adaptive linking coefficient, image denoising, local neighborhood, noise detection