Rectifier Neural Network with a Dual-Pathway Architecture for Image Denoising

Recently deep neural networks based on tanh activation function have shown their impressive power in image denoising. However, much training time is needed because of their very large size. In this letter, we propose a dual-pathway rectifier neural network by combining two rectifier neurons with reversed input and output weights in the same hidden layer. We drive the equivalent activation function and illustrate that it improves the efficiency of capturing information from the noisy data. The experimental results show that our model outperforms other activation functions and achieves state-of-the-art denoising performance, while the network size and the training time are significantly reduced.