Journal of Computer Applications ›› 2018, Vol. 38 ›› Issue (1): 246-254.DOI: 10.11772/j.issn.1001-9081.2017061461

Previous Articles     Next Articles

Improved algorithm of image super resolution based on residual neural network

WANG Yining, QIN Pinle, LI Chuanpeng, CUI Yuhao   

  1. School of Data Science, North University of China, Taiyuan Shanxi 030051, China
  • Received:2017-06-14 Revised:2017-08-10 Online:2018-01-10 Published:2018-01-22
  • Supported by:
    This work is partially supported by the Natural Science Foundation of Shanxi Province (2015011045).

基于残差神经网络的图像超分辨率改进算法

王一宁, 秦品乐, 李传朋, 崔雨豪   

  1. 中北大学 大数据学院, 太原 030051
  • 通讯作者: 秦品乐
  • 作者简介:王一宁(1992-),女,山西长治人,硕士研究生,主要研究方向:深度学习、机器视觉、数字图像处理;秦品乐(1978-),男,山西长治人,副教授,博士,主要研究方向:大数据、机器视觉、三维重建;李传朋(1991-),男,山东济南人,硕士研究生,主要研究方向:深度学习、机器视觉、数字图像处理;崔雨豪(1996-),男,浙江杭州人,主要研究方向:深度学习、数字图像处理。
  • 基金资助:
    山西省自然科学基金资助项目(2015011045)。

Abstract: To efficiently improve the effects of image Super Resolution (SR), a multi-stage cascade residual convolution neural network model was proposed. Firstly, two-stage SR image reconstruction method was used to reconstruct the 2-times SR image and then reconstruct the 4-times SR image; secondly, residual layer and jump layer were used to predict the texture information of the high resolution space in the first and second stages, and deconvolution layer was used to reconstruct 2-times and 4-times SR images. Finally, two multi-task loss functions were constructed respectively by the results of two stages. And the loss of the first stage guided that of the second one, which accelerated the network training and enhanced the supervision and guidance of the network learning. The experimental results show that compared with bilinear algorithm, bicubic algorithm, Super Resolution using Convolutional Neural Network (SRCNN) algorithm and Fast Super Resolution Convolutional Neural Network (FSRCNN) algorithm, the proposed model can better construct the details and texture of images, which avoids the image over smoothing after iterating, and achieves higher Peak Signal-to-Noise Ratio (PSNR) and Mean Structural SIMilarity (MSSIM).

Key words: Super Resolution (SR), Deep Learning (DL), residual block, jump layer, deconvolution, multi-task loss

摘要: 为更有效地提升图像的超分辨率(SR)效果,提出了一种多阶段级联残差卷积神经网络模型。首先,该模型采用了两阶段超分辨率图像重建方法先重建2倍超分辨率图像,再重建4倍超分辨率图像;其次,第一阶段与第二阶段皆使用残差层和跳层结构预测出高分辨率空间的纹理信息,由反卷积层分别重建出2倍与4倍大小的超分辨率图像;最后,以两阶段的结果分别构建多任务损失函数,利用第一阶段的损失指导第二阶段的损失,从而提高网络的训练速度,加强网络学习中的监督指导。实验结果表明,与bilinear算法、bicubic算法、基于卷积神经网络的图像超分辨率(SRCNN)算法和加速的超分辨率卷积神经网络(FSRCNN)算法相比,所提模型能更好地重建出图像的细节和纹理,避免了经过迭代之后造成的图像过度平滑,获得更高的峰值信噪比(PSNR)和平均结构相似度(MSSIM)。

关键词: 超分辨率, 深度学习, 残差块, 跳层, 反卷积, 多任务损失

CLC Number: