臺灣博碩士論文加值系統

分析核磁共振圖像主要是根據各種特徵在原始圖像中提取紋理。然而，大部分的紋理演算法都是屬於要大量計算的方法。灰階等級遊程長度矩陣則是常會看到的架構之一，而且在預處理階段所做的提取功能在計算成本上是非常昂貴的。最近由於圖形處理器技術的蓬勃發展，也成為了用來加速計算過程的理想選擇。在本研究中，我們提供了一個成熟的平行函式庫的範例，該函式庫是用來方便生成灰階等級遊程長度矩陣和同時對多的感興趣區域提取多個特徵。本文所提出的演算法非常容易實現，並且在與中央處理器上執行灰階等級遊程長度矩陣的實驗結果相比，優化的速度比可以增加至5倍以上。

To analysis Magnetic Resonance images, the initial step is to extract textures from the original images according to a variety of features. However, most of the texture extraction algorithms are computational-consuming approaches. GLRLM is the one of common-used architectures and the computational cost of the extraction functions during the pre-processing phase is very expensive. Recently, GPU technology is booming and also the ideal choice to accelerate the current computing process. In this study, we provide an example of a mature parallel primitive that is used to generate a GLRLM and simultaneously extract multiple features for multiple region of interesting in the image. The proposed algorithm is very easy to implement, and the optimized speed ratio can be increased by more than 5 times compared to GLRLM executing on CPU from experimental results.

摘要 i
英文摘要 ii
致謝 iii
目錄 iv
表目錄 vi
圖目錄 vii
一、緒論 1
1.1研究背景 1
1.2研究動機 3
1.3研究目的 4
1.4論文架構 5
二、文獻探討 6
2.1腦部核磁共振影像 6
2.2灰階等級遊程長度矩陣 7
2.3特徵 9
2.4 GPGPU和函式庫 10
2.4.1 CUDA 11
2.4.2 CUB 12
三、研究方法 14
3.1 平行建構GLRLM 14
3.2 平行特徵提取 17
3.3實驗方法 22
3.3.1實驗一 – 平行建構GLRLM結果 22
3.3.2實驗二 – 不同高效能平台之效能差異 22
3.3.3實驗三 – 使用不同精度之效能差異 22
3.3.4實驗四 – 使用不同ROI大小之效能差異 22
四、實驗結果與分析 23
4.1實驗結果 23
4.1.1實驗一 –平行建構GLRLM結果 23
4.1.2實驗二 –不同高效能平台之效能差異 25
4.1.3實驗三 –使用不同精度之效能差異 26
4.1.4實驗四 –使用不同ROI大小之效能差異 27
五、結論與未來展望 28
5.1 結論 28
5.2 研究建議與方向 28
參考文獻 29

Mary M. Galloway. Texture analysis using gray level run lengths. computer graphics and image processing, 4(2):172–179, June 1975. URL https://www.sciencedirect.com/science/article/pii/S0146664X75800086.
Belur V. Dasarathy and Edwin B. Holder. Image characterizations based on joint gray level—run length distributions. Pattern Recognition Letters, 12(8):497–502, August 1991. ISSN 01678655. doi: 10.1016/0167-8655(91) 80014-2. URL http://linkinghub.elsevier.com/retrieve/pii/0167865591800142.
A. Chu, C.M. Sehgal, and J.F. Greenleaf. Use of gray value distribution of run lengths for texture analysis. Pattern Recognition Letters, 11(6):415–419, June 1990. ISSN 01678655. doi: 10.1016/0167-8655(90)90112-F. URL http://linkinghub.elsevier.com/retrieve/pii/016786559090112F.
Nida M. Zaitoun and Musbah J. Aqel. Survey on image segmentation techniques, September 2015. URL https://www.sciencedirect.com/science/article/pii/S1877050915028574.
S. Selvarajah and S. R. Kodituwakku. Analysis and comparison of texture features for content based image retrieval. International Journal of Latest Trends in Computing, 2(1), 2011. URL http://www.ijltc.excelingtech.co.uk/vol2issue1/18-vol2issue1.pdf.
H. H. Loh, J. G. Leu, and R. C. Luo. The analysis of natural textures using run length features. IEEE Transactions on Industrial Electronics, 35(2):323–328, May 1988. ISSN 0278-0046. doi: 10.1109/41.192665.
J. S. Weszka, C. R. Dyer, and A. Rosenfeld. A Comparative Study of Texture Measures for Terrain Classification. IEEE Transactions on Systems, Man, and Cybernetics, SMC-6(4):269–285, April 1976. ISSN 0018-9472. doi: 10.1109/TSMC.1976.5408777.
R.W. Conners and C. A. Harlow. A Theoretical Comparison of Texture Algorithms. IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-2(3):204–222, May 1980. ISSN 0162-8828. doi: 10.1109/TPAMI.1980.4767008.
Robert M. Haralick, Karthikeyan Shanmugam, and others. Textural features for image classification. IEEE Transactions on systems, man, and cybernetics, 3(6):610–621, 1973.
Xiaoou Tang. Texture information in run-length matrices. IEEE Transactions on Image Processing, 7(11):1602–1609, November 1998. ISSN 10577149. doi: 10.1109/83.725367. URL http://ieeexplore.ieee.org/document/725367/.
G. Castellano, L. Bonilha, L. M. Li, and F. Cendes. Texture analysis of medical images. Clinical Radiology, 59(12):1061–1069, December 2004. ISSN 0009-9260, 1365-229X. doi: 10.1016/j.crad.2004.07.008.
S. Herlidou, Y. Rolland, J. Y. Bansard, E. Le Rumeur, and J. D. De Certaines. Comparison of automated and visual texture analysis in MRI: characterization of normal and diseased skeletal muscle. Magnetic resonance imaging, 17(9):1393–1397, 1999.
S. Poonguzhali and G. Ravindran. Automatic classification of focal lesions in ultrasound liver images using combined texture features. Information Technology Journal, 7(1):205–209, 2008. URL http://docsdrive.com/pdfs/ansinet/itj/2008/205-209.pdf.
David Molina, Juli´an P´erez-Beteta, Alicia Mart´ınez- Gonz´alez, Juan Martino, Carlos Vel´asquez, Estanislao Arana, and V´ıctor M. P´erez-Garc´ıa. Influence of gray level and space discretization on brain tumor heterogeneity measures obtained from magnetic resonance images. Computers in Biology and Medicine, 78:49–57, November 2016. ISSN 00104825. doi: 10.1016/j.compbiomed.2016. 09.011. URL http://linkinghub.elsevier.com/retrieve/pii/S0010482516302372.
Karthik Kalyan, Binal Jakhia, Ramachandra Dattatraya Lele, Mukund Joshi, and Abhay Chowdhary. Artificial neural network application in the diagnosis of disease conditions with liver ultrasound images. Advances in bioinformatics, 2014, 2014. URL https://www.hindawi.com/journals/abi/2014/708279/abs/.
Markus Gipp, Guillermo Marcus, Nathalie Harder, Apichat Suratanee, Karl Rohr, Rainer K¨onig, and Reinhard M¨anner. Accelerating the computation of haralick’s texture features using graphics processing units (gpus). In Proceedings of the World Congress on Engineering, volume 1, 2008.
J. Dixon and J. Ding. An empirical study of parallel solutions for GLCM calculation of diffraction images. In 2016 38th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), pages 3969–3972, August 2016. doi: 10.1109/EMBC. 2016.7591596.
Kristina Doycheva, Christian Koch, and Markus K¨onig. Implementing textural features on GPUs for improved real-time pavement distress detection. Journal of RealTime Image Processing, pages 1–12, October 2016. ISSN 1861-8200, 1861-8219. doi: 10.1007/s11554-016-0648-1.
H. Y. Tsai, H. Zhang, C. L. Hung, and G. Min. GPUaccelerated Features Extraction from Magnetic Resonance Images. IEEE Access., PP(99):1–1, 2017. doi: 10.1109/ACCESS.2017.2756624.
Brainweb: Simulated brain database, . URL http://brainweb.bic.mni.mcgill.ca/brainweb/.
Mcbic: About the MRI simulator, . URL http://brainweb.bic.mni.mcgill.ca/brainweb/mri_sim.html.
Cub: Documentation copyright (c) 2011-2016, nvidia corporation. URL http://nvlabs.github.io/cub/index.html.