|本期目录/Table of Contents|

[1]邬凡,杨俊,桂江生.基于注意力编解码器及多残差网络的逆半色调方法[J].浙江理工大学学报,2024,51-52(自科三):369-377.
 WU Fan,YANG Jun,GUI Jiangsheng.An inverse halftoning method based on encoder decoder with attention and multi residual network[J].Journal of Zhejiang Sci-Tech University,2024,51-52(自科三):369-377.
点击复制

基于注意力编解码器及多残差网络的逆半色调方法()
分享到:

浙江理工大学学报[ISSN:1673-3851/CN:33-1338/TS]

卷:
第51-52卷
期数:
2024年自科第三期
页码:
369-377
栏目:
出版日期:
2024-05-10

文章信息/Info

Title:
An inverse halftoning method based on encoder decoder with attention and multi residual network
文章编号:
1673-3851 (2024) 05-0369-09
作者:
邬凡杨俊桂江生
1.浙江理工大学计算机科学与技术学院,杭州 310018;2.嘉兴学院信息科学与工程学院,浙江嘉兴 314001
Author(s):
WU Fan YANG Jun GUI Jiangsheng
1.School of Computer Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China; 2.College of Information Science and Engineering,Jiaxing University, Jiaxing 314001, China
关键词:
逆半色调图像恢复注意力机制编解码器多残差网络清晰度
分类号:
TP391
文献标志码:
A
摘要:
针对当前逆半色调方法恢复的图像存在细节不清晰甚至丢失的问题,提出了一种基于注意力编解码器及多残差网络(Encoder-decoder with attention and multi-residual network, EDAMRNet)的逆半色调方法。首先,设计融合注意力机制的编解码器结构,在其跳跃连接处添加非对称特征融合模块,以有效提取图像上下文信息;然后,构造多残差网络,捕获并保留图像空间细节信息;最后,应用监督注意力模块对图像上下文信息进行加强,再传递到多残差网络,以恢复出高质量的连续色调图像。实验结果表明:该方法与现有最优方法相比,在Urban100和Manga109数据集下的峰值信噪比平均值均提高了0.1 dB,结构相似性平均值分别提高了0.0010和0.0005。该方法能够在提取图像上下文信息的同时保留图像空间细节信息,可更好地恢复图像纹理信息,提高图像清晰度,为图像逆半色调方法研究提供了一种新的方案。

参考文献/References:

1 Guo J M, Sankarasrinivasan S. Digital halftone database (DHD): A comprehensive analysis on halftone types C ]∥ 2018 Asia - Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC). Honolulu, HI, USA. IEEE, 2019: 1091 - 1099.

2]张燕, 张二虎. 图像逆半色调技术研究[J. 西安理工大学学报, 2017, 33(3): 282-289.

3]孔月萍, 曾平, 何波, . LUT Elman 网络相结合的图像逆半调算法[J. 中国图象图形学报, 2007, 12(11): 1988-1991.

4Xiong Z X, Orchard M T, Ramchandran K. Inverse halftoning using waveletsJ. IEEE Transactions on Image Processing, 1999, 8(10): 1479-1483.

5Mese M, Vaidyanathan P P. Look-up table (LUT) method for inverse halftoningJ. IEEE Transactions on Image Processing, 2001, 10(10): 1566-1578.

6Hou X X, Qiu G P. Image companding and inverse halftoning using deep convolutional neural networksEB/OL. (2017-07-01)2023-07-15. https:arxiv.org/abs/170700116.

7Xiao Y, Pan C, Zheng Y, et al. Gradient-guided DCNN for inverse halftoning and image expandingC]∥Asian Conference on Computer Vision. Cham: Springer International Publishing, 2019: 207-222.

8Yuan J, Pan C, Zheng Y, et al. Gradient-guided residual learning for inverse halftoning and image expandingJ. IEEE Access, 2019, 8: 50995-51007.

9Xia M, Wong T T. Deep inverse halftoning via progressively residual learningC]∥Computer Vision-ACCV 2018: 14th Asian Conference on Computer Vision. Perth, Australia. Cham: Springer International Publishing, 2019: 523-539.

10Son C H. Inverse halftoning through structure-aware deep convolutional neural networksJ. Signal Processing, 2020, 173: 107591.

undefined

备注/Memo

备注/Memo:
收稿日期: 2023-07-14
网络出版日期:2023-12-12
基金项目: 浙江省基础公益计划项目(LGG22F020021);嘉兴市科技计划项目(2021AY10071)
作者简介: 邬凡(1999—),男,江西南昌人,硕士研究生,主要从事数字图像处理、深度学习等方面的研究
通信作者: 杨俊,E-mail:juneryoung@zjxu.edu.cn
更新日期/Last Update: 2024-06-19