Ultra HD Video
Fig.1 Image super-resolution results (Column 1: the original images, Column 2-5; images processed by previous excellent super-resolution methods, Column 6: images processed by our method )
In this paper, we present a novel single-image super-resolution method by introducing densely skip connections in a very deep network. In the proposed network, the feature maps of each layer are propagated into all subsequent layers, providing an effective way to combine the low-level features and high-level features to boost the reconstruction performance. In addition, the densely skip connections enable short paths to be built directly from the output to each layer, alleviating the vanishing-gradient problem of very deep networks. Moreover, the proposed method substantially reduces the number of parameters, enhancing the computational efficiency. Further, deconvolution layers are integrated into the network to learn the upsampling filters and to speedup the reconstruction process. We evaluate the proposed method using images from four benchmark datasets and set a new state of the art.