JPEG 由國際標準組織(International Standardization Organization，簡稱ISO) 和國際電話電報諮詢委員會(International Telegraph and Telephone Consultative Committee ，簡稱CCITT) 所建立的一個數位影像壓縮標準，主要是用於靜態影像壓縮方面。JPEG 採用可失真(Lossy) 編碼法的概念，利用數位餘弦轉換法(Discrete Cosine Transform，簡稱DCT) 和量化器將影像資料中較不重要的部份去除，僅保留重要的資訊，以達到高壓縮率的目的。雖然被JPEG 處理後的影像會有失真的現象，但由於JPEG 的失真比例可以利用參數來加以控制；一般而言，當壓縮率( 即壓縮過後的大小除以原有資料量的結果) 在5% ～15% 之間時，JPEG 依然能保証其適當的影像品質，這是一般無失真壓縮法所作不到的。
JPEG 影像格式在一般生活上，例如網際網路或數位相機，是非常常見的。藉由壓縮可以縮減資料量的大小，以方便我們儲存或傳輸。本論文提出一個結合資料存取架構的演算法，在壓縮影像的同時利用頻率域之資訊得到每筆壓縮資料的重要度，當儲存的資料量不多的時候，所有儲存的資料可以趨近於無失真壓縮，隨著儲存資料的增加，演算法自動依據每筆資料的重要度來調整儲存空間的分配，同時根據Rate-Distortion的理論使所有儲存的影像之間的品質更平滑且壓縮效果可以更好。最後我們與ㄧ般所使用的JPEG壓縮標準來做比較，其影像之間的平滑度及壓縮效果都有顯著的改善。

JPEG (Joint Photographic Experts Group) is an ISO/IEC group of experts that develops and maintains standards for a suite of compression algorithms for computer image files. It is now regarded as a digital image compression standard of the International Standardization Organization (ISO) and the International Telegraph and Telephone Consultative Committee (CCITT). JPEG is mainly used in static image compression. It is a standard of lossy coding that uses discrete cosine transform (DCT) and quantizer to remove unimportant part, to retain important information, and to achieve the goal of high compression rate. Although the image will be distorted after JPEG processing, the distortion can be controlled by some parameters. Generally speaking, when you compress 5% ~ 15% off an original image, JPEG algorithm can still guarantee its suitable image quality, and this is some of the general lossless compression standard can not achieve.

The JPEG image format is commonly used in daily life, for example in the internet web pages or in digital cameras. Although compression reduces the quality of images, it enables them to be stored or transmitted easier. This thesis proposes an algorithm that adaptively adjusts compression parameters according to data storage utilization. The algorithm acquires importance of every compressed data from frequency domain. When fewer images are stored, all the compressed data are store and the quality of all stored images is close to lossless compression. When the number of images increases, the algorithm dynamically adjusts storage space assignment according to the importance of each data. Meanwhile, it can smooth down the variation of quality among images and obtains better compression efficiency according to the rate-distortion theory. The experiments show that the quality variation and the compression efficiency are obviously improved when compared with standard JPEG compression.

[1] J.M. Shapiro, “Embedded Image Coding Using Zerotrees of Wavelet Coefficients”, IEEE Transactions on Signal Processing, Sp-41:3445-3462, December 1993
[2] S.P. Lloyd, “Least Squares Quantization in PCM”, IEEE Transactions on Information Theory, IT-28:127-135, March 1982
[3] Feng Pan, “Adaptive Image Compression using Local Pattern Information”, Pattern Recognition Letters, Volume 23, Issue 14, December 2002
[4] Pennebaker, William B. /Mitchell, Joan L. /Van Nostrand Reinhold. JPEG still image data compression standard. New York: Van Nostrand Reinhold, c1993
[5] G.K. Wallace, “The JPEG Still Picture Compression Standard”, IEEE Transaction on Consumer Electronics, Feb. 1992
[6] Rafael C. Gonzalez. Digital Image Processing, 2/E. Prentice Hall 2002
[7] James F. Blinn, “What’s the Deal with the DCT?” California Institute of Technology, Jim Blinn’s Corner
[8] C.E. Shannon, “A mathematical theory of communication”, Bell Syst. Tech. Journal, vol. 27, pp. 379-423, 1948
[9] S. Arimoto, “An Algorithm for Computing the Capacity of Arbitrary Discrete Memoryless Channels”, IEEE Transactions on Information Theory, IT-18:14-20, January 1972
[10] R.E. Blahut, “Computation of Channel Capacity and Rate Distortion Function”, IEEE Transactions on Information Theory, IT-18:460-473, July 1972
[11] Khalid Sayood, “Introduction to Data Compression”, 2/E. Morgan Kaufmann 2000
[12] Zhihai He and Sanjit K. Mitra, “From Rate-Distortion Analysis to Resource-Distortion Analysis”, IEEE Circuits and Systems Magazine, Third Quarter 2005
[13] Antonio Ortega and Kannan Ramchandran, “Rate-Distortion Methods for Image and Video Compression”, IEEE Signal Processing Magazine, November 1998
[14] Jaehan In, Shahram Shirani, and Faouzi Kossentini, “On RD Optimized Progressive Coding Using JPEG”, IEEE Transactions On Image Processing, VOL. 8, NO.11, November 1999
[15] T. H. Cormen, C.E. Leiserson, and R. L. Rivest, “Introduction to Algorithms”, MIT Press, 1990
[16] Robert Sedgewick, “Algorithms in C”, Addison-Wesley, 1998
[17] van der Vleuten, R.J, “Flexible Storage of Images with Application to Digital Cameras”, IEEE Conference, Volume 2, 7-10 Oct. 2001
[18] Cormac Herley. “Storage of Digital Camera Images”, In Int. Conf. Image Proc. (ICIP’99), Kobe, Japan, Oct.24-28, 1999, vol. 3, pp. 940-942
[19] Furuie, S.S, “Flexible storage architecture for large PACS”, IEEE Conference. 22-25 Sept. 2002
[20] The USC-SIPI Image Database. http//sipi.usc.edu/database/
[21] ftp://ftp.simtel.net/pub/simtelnet/msdos/graphics/jpegsr6.zip