[1]陶霖密,袁春,王浩达.一种恢复图像自然色彩的重构方法[J].智能系统学报,2019,14(5):877-881.[doi:10.11992/tis.201805011]
TAO Linmi,YUAN Chun,WANG Haoda.Reconstruction method for restoring the natural color of an image[J].CAAI Transactions on Intelligent Systems,2019,14(5):877-881.[doi:10.11992/tis.201805011]
点击复制
TAO Linmi,YUAN Chun,WANG Haoda.Reconstruction method for restoring the natural color of an image[J].CAAI Transactions on Intelligent Systems,2019,14(5):877-881.[doi:10.11992/tis.201805011]
一种恢复图像自然色彩的重构方法
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
14
期数:
2019年第5期
页码:
877-881
栏目:
学术论文—机器学习
出版日期:
2019-09-05
- Title:
- Reconstruction method for restoring the natural color of an image
- Keywords:
- color filter array; compressed sensing; sparse coding; dictionary learning; image reconstruction; natural color
- 分类号:
- TP18
- 摘要:
- 现代数码相机是通过颜色过滤矩阵在每个像素位置采集一个颜色分量，重构出全彩色数字图像。压缩感知理论证明了该重构是误差有界的，但在实际应用时却隐含着一个问题：重构图像所需的稀疏编码字典是从图像数据库学习出来的，而目前数字图像都是重构出来的，因此存在着从重构的图像学习字典去重构图像的循环悖论。针对这个问题，提出并构建了新的完全采样彩色图像的Sandwich图像数据集，打破了压缩感知理论在应用于图像重构时的循环悖论，使得压缩感知方法能够真正地重建自然彩色图像。Sandwich图像数据集的构建及其训练得到的字典可以应用于如图像超分辨率重构、去噪、修复等领域。深入的图像重建实验表明，使用sandwich图像集训练的字典不论是字典原子特性还是由其重构得到的图像质量均好于基于传统数据集的结果。
- Abstract:
- Currently, the digital camera captures one color component at each pixel location through a color filter array and reconstructs a full-color digital image. The compressed sensing theory has proven that this reconstruction is based on the error bound with sparse coding dictionary. However, in practice, there is a hidden problem:the sparse coding dictionary needed for image reconstruction is learned from the image database while the current digital images are reconstructed. Thus, there is a cyclic paradox of reconstructing images from the reconstructed image learning dictionary. To solve this problem, this study proposes and constructs a fully sampled color image dataset called Sandwich to break the cyclic paradox of compressed sensing theory in the application of image reconstruction, enabling the compressed sensing method to reconstruct truly natural color images. The dictionary trained from the constructed Sandwich dataset can be applied to several domains, such as image super-resolution reconstruction, denoising, and restoration. The in-depth image reconstruction experiments show that the dictionary trained using the Sandwich image dataset is better than the dictionary trained using the traditional dataset in terms of both dictionary atomic characteristics and the reconstructed image quality.
- 参考文献/References:
-
[1] KIMMEL R. Demosaicing:image reconstruction from color CCD samples[J]. IEEE transactions on image processing, 1999, 8(9):1221-1228.
[2] DONOHO D L. Compressed sensing[J]. IEEE transactions on information theory, 2006, 52(4):1289-1306.
[3] GHARBI M, CHAURASIA G, PARIS S, et al. Deep joint demosaicking and denoising[J]. ACM transactions on graphics, 2016, 35(6):191.
[4] ZHANG Meiqing, TAO Linmi. A patch aware multiple dictionary framework for demosaicing[C]//Proceedings of the 12th Asian Conference on Computer Vision. Singapore, Singapore, 2014:236-251.
[5] ZHANG Lei, WU Xiaolin, BUADES A, et al. Color demosaicking by local directional interpolation and nonlocal adaptive thresholding[J]. Journal of electronic imaging, 2011, 20(2):023016.
[6] MAIRAL J, ELAD M, SAPIRO G. Sparse representation for color image restoration[J]. IEEE transactions on image processing, 2008, 17(1):53-69.
[7] MOGHADAM A A, AGHAGOLZADEH M, KUMAR M, et al. Compressive framework for demosaicing of natural images[J]. IEEE transactions on image processing, 2013, 22(6):2356-2371.
[8] FRANZEN R. Kodak lossless true color image suite[EB/OL]. (1999-11-15)[2018-03-22]. http://r0k.us/graphics/kodak.
[9] GUNTURK B K, GLOTZBACH J, ALTUNBASAK Y, et al. Demosaicking:color filter array interpolation[J]. IEEE signal processing magazine, 2005, 22(1):44-54.
[10] GUNTURK B K, ALTUNBASAK Y, MERSEREAU R M. Color plane interpolation using alternating projections[J]. IEEE transactions on image processing, 2002, 11(9):997-1013.
[11] LI Xin, GUNTURK B, ZHANG Lei. Image demosaicing:a systematic survey[C]//Proceedings of Visual Communications and Image Processing 2008. San Jose, United States, 2008:68221J.
[12] NAKAMURA J. Image sensors and signal processing for digital still cameras[M]. Boca Raton:CRC Press, 2005.
[13] LOSSON O, MACAIRE L, YANG Y. Comparison of color demosaicing methods[J]. Advances in imaging and electron physics, 2010, 162:173-265.
[14] CANDèS E J, ROMBERG J, TAO T. Robust uncertainty principles:exact signal reconstruction from highly incomplete frequency information[J]. IEEE transactions on information theory, 2006, 52(2):489-509.
- 相似文献/References:
-
[1]赵春晖,刘巍.压缩感知理论及其在成像技术中的应用[J].智能系统学报,2012,7(1):21.
ZHAO Chunhui,LIU Wei.Compressive sensing theory and its application in imaging technology[J].CAAI Transactions on Intelligent Systems,2012,7():21.
[2]郑晓,薄华,孙强.QR分解与特征值优化观测矩阵的算法研究[J].智能系统学报,2015,10(1):149.[doi:10.3969/j.issn.2013-0934.201309034]
ZHENG Xiao,BO Hua,SUN Qiang.An algorithm for measurement matrix based on QR decomposition and eigenvalue optimizatio[J].CAAI Transactions on Intelligent Systems,2015,10():149.[doi:10.3969/j.issn.2013-0934.201309034]
[3]唐琳,焦淑红.基于近似消息传递算法的压缩感知雷达成像方法[J].智能系统学报,2015,10(4):592.[doi:10.3969/j.issn.1673-4785.201411025]
TANG Lin,JIAO Shuhong.Compressed sensing radar imaging using approximate message passing[J].CAAI Transactions on Intelligent Systems,2015,10():592.[doi:10.3969/j.issn.1673-4785.201411025]
[4]史东承,倪康.压缩感知W-HOG特征的运动手势跟踪[J].智能系统学报,2016,11(1):124.[doi:10.11992/tis.201507005]
SHI Dongcheng,NI Kang.Motion gesture tracking based on compressed sensing W-HOG features[J].CAAI Transactions on Intelligent Systems,2016,11():124.[doi:10.11992/tis.201507005]
[5]张龙,陈宸,韩宁,等.压缩感知理论中的建筑电气系统故障诊断[J].智能系统学报,2014,9(2):204.[doi:10.3969/j.issn.1673-4785.201310026]
ZHANG Long,CHEN Chen,HAN Ning,et al.Fault diagnosis of electrical systems in buildingsbased on compressed sensing[J].CAAI Transactions on Intelligent Systems,2014,9():204.[doi:10.3969/j.issn.1673-4785.201310026]
[6]李霞丽,吴立成,樊艳明.易于硬件实现的压缩感知观测矩阵的研究与构造[J].智能系统学报,2017,12(3):279.[doi:10.11992/tis.201606037]
LI Xiali,WU Licheng,FAN Yanming.Study and construction of a compressed sensing measurement matrix that is easy to implement in hardware[J].CAAI Transactions on Intelligent Systems,2017,12():279.[doi:10.11992/tis.201606037]
[7]孙必慎,石武祯,姜峰.计算视觉核心问题：自然图像先验建模研究综述[J].智能系统学报,2019,14(1):71.[doi:10.11992/tis.201804019]
SUN Bishen,SHI Wuzhen,JIANG Feng.Core problem in computer vision: survey of natural image prior models[J].CAAI Transactions on Intelligent Systems,2019,14():71.[doi:10.11992/tis.201804019]
[8]邵凯,闫力力,王光宇.压缩感知重构算法的两步深度展开策略研究[J].智能系统学报,2023,18(5):1117.[doi:10.11992/tis.202204029]
SHAO Kai,YAN Lili,WANG Guangyu.Two-step deep unfolding strategy for compressed sensing reconstruction algorithms[J].CAAI Transactions on Intelligent Systems,2023,18():1117.[doi:10.11992/tis.202204029]
备注/Memo
收稿日期:2018-05-09。
基金项目:国家自然科学基金项目（61672017）.
作者简介:陶霖密,男,1962年生,副教授,主要研究方向为人机交互、计算机视觉与模式识别。承担国家重点基金情感计算项目,以及与IBM、INTEL、SI-EMENS的国际合作基金项目等。发表学术论文多篇;袁春,男,1969年生,副研究员,博士生导师,主要研究方向为机器学习、计算机视觉、视频分析与处理。先后负责和参与微软、欧盟、国家自然科学基金、863和973等重要研究课题。取得专利15项,发表学术论文70余篇王浩达,男,1992年生,硕士研究生,主要研究方向为机器学习、计算机视觉。
通讯作者:陶霖密.E-mail:linmi@tsinghua.edu.cn
更新日期/Last Update:
1900-01-01