臺灣博碩士論文加值系統

　　在影像壓縮領域中，LBG是一個速度快，結構簡單易懂的方法，而訓練後之壓縮品質亦可，在向量量化(Vector Quantization, VQ)領域中佔有重要一席之地。另一種分割式的演算法細胞分裂演算法，結構更為簡單，省去LBG的檢查收歛情形，以一分裂為二的方式加減向量值，速度更快，但是PSNR值卻不盡理想。
　　本論文提出一個基於細胞分裂法的理念上更快速的分裂向量的演算方法，同時要避免因為缺代優化訓練而造成壓縮品質不夠理想的遺憾，所以也參酌了LBG的方法應用於其中，期使PSNR值提高於水準之上，欲使造就出一個結構簡單、訓練品質優良、壓縮時間快速的演算法。

In the field regarding image compression, as LBG is a fast and easily understanding method with simple construction, and its compressed quality after training is acceptable, LBG is always considered an important technique in VQ(Vector Quantization)field. The other divisive algorithm is Cell Divisive Algorithm. Its construction is much simpler as it removes the precedure of checking convergent status in LBG, and add or subtract vector's value after dividing one into two. The whole procedure is faster, but the PSNR value is not satisfactory.
The thesis is to propose a faster divisive vector algorithm using cell divisive algorithm. Since the defect of disqualified compression quality should be avoided due to lack of optimized training, LBG method is also applied in the algorithm to improve PSNR value. The purpose of the algorithm is to set up a easily-constructed, quickly-compressed algorithm with well-trained quality.

第一章　緒論
1-1 研究背景................1
1-2 研究動機與目的...........2
1-3 研究流程................4
1-4 論文架構................5
1-5 研究範圍與限制...........6
第二章 文獻探討
2-1 影像基本概念.............7
2-2 BMP檔原理...............8
2-3 向量量化(VQ)............8
2-4 SOM演算法..............10
2-5 HSOM演算法.............13
2-6 LBG演算法..............15
2-7 細胞分裂演算法..........16
2-8 LazySOM演算法..........18
2-9 影像壓縮評估............21
第三章 研究方法
3-1 方法源起................23
3-2 十字擴散演算法理論與流程..24
第四章 實驗結果
4-1 實驗環境................31
4-2 Airplane圖的實驗........33
4-3 Lena圖的實驗............38
4-4 Baboon圖的實驗..........43
4-5 Aerial圖的實驗..........48
4-6 綜合分析................52
第五章 結論、心得與未來研究
5-1 結論....................54
5-2 後續可進行之研究.........61

[1]陳湘揚，網路概論，博碩文化股份有限公司，2009。
[2]林東清，知識管理，智勝文化事業有限公司，2009。
[3]蔡文輝、林家楨、張真誠，資料壓縮技術與應用，全華圖書股份有限公司，2007。
[4]鄭柏左，色彩理論與數位影像，新文京出版社，2004。
[5]林于峻，具新的高效能與高效率之增強型自組織特徵映射圖於影像壓縮問題之研發，國立屏東科技大學資訊管理學系碩士論文，2009。
[6]Gray, R.M., "Vector Quantization," IEEE ASSP Magazine, Vol. No. 2, pp. 4-29, 1984.
[7]Kohonen, T., "Self-organizing Maps," Springer-Verlag, Berlin, 1995.
[8]Kohonen, T., "The Self-Organizing Maps," In:Proceedings of the IEEE, Vol. 78, No. 9, 1464-1480, 1990.
[9] Madeiro, F., Vilar, R. and Neto, B., “A Self-Organizing Algorithm for Image Compression,” IEEE Trans. on Neural Networks, Vol. 28, pp. 146-150, 1998.
[10] Barbalho, M., Duarte, A., Neto, D., Costa, F. and Netto, A., “Hierarchical SOM Applied to Image Compression,” In: Proceedings of International Joint Conference on Neural Networks, Vol. 1, pp. 35-48, 2002.
[11] Su, M.C., Chang, H.T. “Fast self-organizing feature map algorithm,” IEEE Transaction on Neural Networks, Vol. 13, No. 3, pp. 721-733, 2000.
[12] Linde, Y., Buzo, A. and Gray, R., “An Algorithm for Vector Quantizer Design,” IEEE Trans. on Communications, Vol. COM-28, pp. 84-95, 1980.
[13] Y. C. Hu and C. C. Chang, “A Progressive Codebook Training Algorithm for Image Quantization,” Proceedings of the Fifth Asia-Pacific Conference on Communications and Fourth Optoelectronics and Communications Conference (APCC/OECC’99), Beijing, China, Vol. 2. pp. 936-939, 1999.
[14] Lehmann, E.L. and Casella, George, “Theory of Point Estimation,” 2nd Edition, Springer-Verlag, New York, 1998.