关键词: 纹理图像, 图像增强, 小波变换, 脉冲耦合神经网络, 中值滤波

Abstract: To solve the restriction problem of direction for the detail enhancement of texture image, this paper proposes an image enhancement algorithm which merges wavelet transform and improved Pulse Coupled Neural Network（PCNN）. Firstly, the image is decomposed by two-dimension discrete wavelet transform to extract the high-frequency component diagram. Then, it improves pulse coupled neural network which uses the local gradients value of image pixel as the link intensity coefficient and adds the lateral inhibition signal to dynamic threshold function, meanwhile, the high-frequency component diagram is enhanced by the improved pulse coupled neural network. Finally, the reconstructed image by inverse wavelet transform is smoothed nonlinearly by median filtering and enhancement for details of texture image is implemented. The experimental results show that this algorithm can effectively reduce the restriction problem of direction for detail enhancement of image. The details of texture image are more plentiful and the overall contrast has a certain improvement after the texture image is enhanced by this algorithm.

Key words: texture image, image enhancement, wavelet transform, Pulse Coupled Neural Network（PCNN）, median filtering