Image super-resolution with densely connected convolutional networks

摘要

This paper proposes a new model for single image super-resolution (SR) task by utilizing the design of densely connected convolutional networks (DenseNet). The proposed method is an end-to-end model which is able to learn mapping between low- and high-resolution images. The proposed method takes the low-resolution images as input and generates its high-resolution version. Unlike those conventional methods which adjust each component of convolutional networks separately, our model jointly optimizes all layers. Besides, the proposed model has a lightweight structure and is extensively evaluated on widely adopted data sets. In our experiments, the proposed method outperforms state-of-the-art methods both qualitatively and quantitatively. In addition, we also carried out experiments in terms of different designs and configurations to achieve better balance between reconstruction performance and speed in this paper.