《计算机应用研究》|Application Research of Computers

学习小波超分辨率系数的人脸超分算法

Wavelet-based deep learning algorithm for face super-resolution

免费全文下载 (已被下载 次)  
获取PDF全文
作者 刘颖,孙定华,公衍超
机构 西安邮电大学 图像与信息处理研究所,西安 710121
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)12-065-3830-06
DOI 10.19734/j.issn.1001-3695.2019.08.0570
摘要 目前基于卷积神经网络的超分方法虽然在峰值信噪比和结构相似性评价指标上能取得优异的结果,但是得到的超分图像视觉质量较差,会丢失人脸五官区域的细节信息。针对这一现象,设计了一种新的深度神经网络来预测超分小波系数以获得信息丰富的超分辨率人脸图像。首先利用人脸图像的先验知识手动地给予五官区域更多的关注,然后在网络中引入线性低秩卷积运算,最后利用长距离依赖的思想补充超分图像的细节。实验验证该算法可以在获得较高的峰值信噪比和结构相似性的同时,使超分人脸图像五官区域更加清晰、视觉质量更优。
关键词 深度学习; 卷积神经网络; 小波; 超分辨率重建
基金项目 国家自然科学基金资助项目(61801381)
陕西省国际合作交流项目(2018KW-003)
本文URL http://www.arocmag.com/article/01-2020-12-065.html
英文标题 Wavelet-based deep learning algorithm for face super-resolution
作者英文名 Liu Ying, Sun Dinghua, Gong Yanchao
机构英文名 Center for Image & Information Processing,Xi'an University of Posts & Telecommunications,Xi'an 710121,China
英文摘要 At present, super-resolution method based on CNN can obtain excellent results in the evaluation metric of peak signal-to-noise ratio and the structural similarity index, but the visual perceptual quality of super- resolution image is poor, and the details of the facial features are lost. In order to solve these problem, this paper designed a new deep neural network to predict the super-resolution wavelet coefficients to get clear super-resolution face images. Firstly, it used the prior knowledge of face images to manually give the facial features more attention. Then it introduced linear low-rank convolution in the network, and finally used the idea of long-distance dependence to supplement the details of the super-resolution images. The experimental results show that the method in this paper has achieved competitive results in the evaluation metric of peak signal-to-noise ratio and structural similarity index, and the visual perceptual quality of its super-resolution face images are excellent.
英文关键词 deep learning; convolutional neural network(CNN); wavelet; super resolution(SR)
参考文献 查看稿件参考文献
 
收稿日期 2019/8/3
修回日期 2019/9/27
页码 3830-3835
中图分类号 TP391.41
文献标志码 A