An Image Super-resolution Reconstruction Method by Using of Deep Learning
Defu Qiu, Lixin Zheng, Shengxiang Zhang, Ying Liu · 2019
In order to better apply clothing style design in clothing industries, this paper presents an image super-resolution reconstruction method based on deep learning. In this proposed method, we add a sub-pixel convolutional layer and a replacement concatenated convolution kernel, which makes a more efficient super-resolution. Therefore, a convolutional neural network consisting of a suitable depth is designed to ensure the quality of the image reconstruction, and cascaded small convolution kernels is applied to improve the running speed. The experimental results show that compared with the Efficent Sub-Pixel Convolutional Neural Network (ESPCN) algorithm, the reconstruction quality of the high-resolution image reconstructed PSNR by the algorithm is improved by 2.57dB and the reconstruction speed by 68 milliseconds.