Multi-level Residual Attention Network for Speckle Suppression.

In order to achieve effective speckle suppression, we propose a multi-level residual attention network by combining with multi-level block and residual channel attention network, which is suitable for speckle suppression. Firstly, the network model performs a simple shallow feature extraction for the input noise image through two convolution layers. Then, the residual attention network is used to extract the deep features. Finally, a convolution layer and residual learning are used to generate the final denoised image. Experimental results show that the proposed method can effectively suppress the noise and preserve the edge details of the image.