基于多原子快速匹配追踪的图像编码算法

邓承志; 汪胜前; 曹汉强

doi:10.3724/SP.J.1146.2008.01150

基于多原子快速匹配追踪的图像编码算法

doi: 10.3724/SP.J.1146.2008.01150

邓承志^{①;汪胜前},
汪胜前,
曹汉强

基金项目:

国家自然科学基金(60462003，60772091)资助课题

计量
- 文章访问数: 3542
- HTML全文浏览量: 104
- PDF下载量: 945
- 被引次数: 0
出版历程
- 收稿日期: 2008-09-16
- 修回日期: 2009-01-05
- 刊出日期: 2009-08-19

Multi-atoms Rapid Matching Pursuit-Based Image Coding Algorithm

摘要

摘要: 该文提出一种多原子快速匹配追踪信号稀疏分解算法，并将其应用于静态图像编码。多原子匹配追踪通过每次迭代选取多个原子的形式，实现信号的快速稀疏分解。在此基础上，通过构造多尺度脊波字典实现图像的稀疏分解，并对稀疏分解的数据进行自适应量化和编码。实验结果表明，多原子匹配追踪获得了与匹配追踪相当的逼近性能，同时极大地提高了稀疏分解的速度。新的编码算法在低比特率情况下，获得了比JPEG2000更理想的编码性能。
- 图像编码;稀疏表示;冗余字典;匹配追踪;多尺度脊波字典
Abstract: In this paper, a Multi-Atoms rapid Matching Pursuit (MAMP) algorithm for signal sparse decomposition and its application to image coding are proposed. The MAMP algorithm decompose signal sparsely by selection several atoms at each iteration. A multiscale ridgelet dictionary is constructed and used to represent image based MAMP. The sparsely decomposed data are adaptively quantized and encoded. Experimental results show that the approximation performances of the MAMP algorithm are comparable with those of the matching pursuit. Meanwhile, the computation speed is greatly improved. On the other hand, the performances of the new coding scheme are shown to compare favorably against those of the state of the art JPEG2000 scheme at low bit rate.

HTML全文

参考文献(1)

Taubman D and Marcellin M. JPEG-2000: ImageCompression Fundantmentals, Standards and Practice [M].Massachusetts: Kluwer Academic Publishers, 2001: 212-379.[2]Mallat S and Zhang Z. Matching pursuit with time-frequencydictionaries [J].IEEE Transactions on Signal Processing.1993, 41(12):3397-3415[3]Gribonval R and Vandergheynst P. On the exponentialconvergence of matching pursuits in quasi-incoherentdictionaries [J].IEEE Transactions on Information Theory.2006, 52(1):255-261[4]Peotta L and Vandergheynst P. Matching pursuit with blockincoherent dictionaries [J].IEEE Transactions on SignalProcessing.2007, 55(9):4549-4557[5]Aharon M and Elad M. Sparse and redundant modeling ofimage content using an image-signature-dictionary [J].SIAMJournal on Imaging Sciences.2008, 1(3):228-247[6]Mairal J, Sapiro G, and Elad M. Learning multiscale sparserepresentations for image and video restoration [J].SIAMMultiscale Modeling and Simulation.2008, 7(1):214-241[7]Mairal J, Elad M, and Sapiro G. Sparse representation forcolor image restoration [J]. IEEE Transactions on ImageProcessing, 2008, 17(1): 53-69.[8]Ventura R M Figueras I, Vandergheynst P, and Frossard P.Low-rate and flexible image coding with redundantrepresentation [J].IEEE Transactions on Image Precessing.2006, 15(3):726-739[9]Gribonval R. Fast matching pursuit with a multiscaledictionary of gaussian chirps [J].IEEE Transactions onSignal Processing.2001, 49(5):994-1001[10]De Vleeschwouwer C and Macq B. Subband dictionaries forlow-cost matching pursuit of video residues [J].IEEETransactions on Circuits and Systems for Video Technology.1999, 9(7):984-993[11]Schmid-Saugeon P and Zakhor A. Dictionary design formatching pursuit and application to motion-compensatedvideo coding [J].IEEE Transactions on Circuits and Systemsfor Video Technology.2004, 14(6):880-886[12]Chou Y T, Hwang W L, and Huang C L. Gain-shapeoptimized dictionary for matching pursuit video coding [J].Signal Processing.2003, 83(9):1937-1943[13]Jost P, Vandergheynst P, and Frossard P. Tree-based pursuit:algorithm and properties [J].IEEE Transactions on SignalProcessing.2006, 54(12):4685-4697[14]Candes E J. Harmonic analysis of neural networks [J].Applied and Computational Harmonic Analysis.1999, 6(2):197-218[15]Shoa A and Shirani S. Adaptive quantization for matchingpursuit [C]. IEEE International Workshop on MultimediaSignal Processing, Victoria, Canada, Oct. 3-6, 2006: 71-74.

施引文献

资源附件(0)

访问统计