A Hybrid CNN for Image Denoising

Deep convolutional neural networks (CNNs) with strong learning abilities have been used in the field of image super-resolution. However, some CNNs depends on a single deep network to training an super-resolution model, which will poor performance complex screens. To address this problem, we propose hybrid denoising CNN (HDCNN). HDCNN is composed dilated block (DB), RepVGG (RVB) and feature refinement (FB), convolution. DB combines convolution, batch normalization (BN), common convolutions, activation function ReLU obtain more context information. RVB uses parallel combination convolution BN, extract complementary width features. FB accurate information via refining obtained from RVB. A collaborates residual operation construct clean image. These key components make good denoising. Experiment shows that proposed enjoys effect public datasets.