Page 311 - 《软件学报》2020年第12期

刘宇男 等:基于级联密集网络的轮廓波变换域图像复原 3977

Fig.10 Qualitative results of different methods (scale factor is 4)
图 10 不同方法的定性结果(缩放因子为 4)

4 总 结
本文提出了一种结合级联密集型神经网络和轮廓波变换的图像复原方法,该方法通过连接更加紧密的卷
积神经网络,充分地挖掘和利用了不同层次的图像特征,并且有效地解决了深层网络中存在的长期依赖问题.在
此基础上,本文引入轮廓波变换,提高了结构及纹理等高频细节的复原能力.标准测试集上的测试结果表明:本
文方法在图像去噪、JPEG 解压缩以及超分辨率这 3 个复原任务上均展现了最优的性能,不但获得了更好的客
观评估结果,而且主观的复原图像中包含了更加逼真的结构及纹理细节.

References:
[1] Zhang K, Zuo WM, Zhang L. Learning a single convolutional super-resolution network for multiple degradations. In: Proc. of the
IEEE Int’l Conf. on Computer Vision and Pattern Recognition (CVPR). 2018. 3262−3271. [doi: 10.1109/CVPR.2018.00344]
[2] Zhang K, Zuo WM, Chen YJ, Meng DY, Zhang L. Beyond a Gaussian denoiser: Residual learning of deep CNN for image
denoising. IEEE Trans. on Image Processing (TIP), 2017,26(7):3142−3155. [doi: 10.1109/TIP.2017.2662206]
[3] Dabov K, Foi A, Katkovnik V, Egiazarian KO. Image denoising by sparse 3-d transform-domain collaborative filtering. IEEE Trans.
on Image Processing (TIP), 2007,16(8):2080−2095. [doi: 10.1109/TIP.2007.901238]
[4] Xu J, Zhang L, Zuo WM, Zhang D, Feng XC. Patch group based nonlocal self-similarity prior learning for image denoising. In:
Proc. of the IEEE Int’l Conf. on Computer Vision (ICCV). 2015. 244−252. [doi: 10.1109/ICCV.2015.36]
[5] Wang XH, Zhu YH, Lv F, Su X, Song CM. Cauchy distribution NSST-HMT model and its applications in image denoising.
Chinese Journal of Computers, 2018,41(11):2496−2508 (in Chinese with English abstract). http://cjc.ict.ac.cn/qwjs/view.asp?id=
5087 [doi: 10.11897/SP.J.1016.2018.02496]
[6] Zhou F, Yang WM, Liao QM. Interpolation-Based image super-resolution using multisurface fitting. IEEE Trans. on Image
Processing (TIP), 2012,21(7):3312−3318. [doi: 10.1109/TIP.2012.2189576]
[7] Chantas GK, Galatsanos NP, Woods NA. Super-Resolution based on fast registration and maximum a posteriori reconstruction.
IEEE Trans. on Image Processing (TIP), 2007,16(7):1821−1830. [doi: 10.1109/TIP.2007.896664]
[8] Li M, Cheng J, Le X, Luo HM. Super-Resolution based on sparse dictionary coding. Ruan Jian Xue Bao/Journal of Software, 2013,
23(5):1315−1324 (in Chinese with English abstract). http://www.jos.org.cn/1000-9825/3989.htm [doi: 10.3724/SP.J.1001.2012.
03989]