臺灣博碩士論文加值系統

目的：本研究主要目的是在現有的內視鏡系統中加入立體視覺的技術，希望透過影像處理的技術，讓操作者及實習者或助手皆能同步看到立體的影像，而其他的醫療人員可藉原來的平面顯示器看手術的進行。因為是立體視覺的影像，操作者對標的物在「距離」上的拿捏會比較精確，手術儀器的操作上能更快，手術時間便可以更短。

方法：我們實驗所使用的內視鏡影像是由台北榮民總醫院所提供，要產生立體視覺有兩個部分。首先，從輸入的影像中獲取深度資訊，我們利用Matlab（R2009a）來做模擬並驗證。第二，我們使用兩個FPGA（Altera的DE -2 70[26]）作為硬件開發平台，然後將位移演算法與之前所提出的影像深度資訊相結合，輸入端為AV信號而輸出端為VGA信號。處理後的影像將輸出到HMD（頭戴式顯示器CybermanTM），因為有兩個VGA輸出，因此每隻眼睛接收到不同的圖像從FPGA平台。

結果與結論：以Matlab模擬實際腸道鏡的影像訊號，經過FPGA的處理後，我們證明了在內視鏡環境下以反光強度作為深度資訊的可行性。本研究依據Matlab轉化後的結果作為硬體上雙眼視差演算法的依據，將原始影像通過FPGA做影像處理後的結果還有許多可以優化的空間。

Objective：
The main purpose of this study is to add the technique of stereo-vision to the traditional endoscopic display device. By which, the operators and the assistants can view real-time 3D video image that can help them to have a better sense of depth during a surgical operation. In the mean time the other peoples in the surgical team can watch the operation from a traditional flat panel display (LCD). Due to having the sense of depth, the surgery can be operated more precisely and the operation time can be shortened.

Method：
Our testing video is provided by Taipei Veterans General Hospital and then we take some video fragment for verification. In order to generate a stereo image, there are two parts. First, acquiring depth-data from the given image, we simulated and demonstrated it by Matlab(R2009a). Second, we are using two FPGA (Altera DE-2 70[2]) as a hardware platform for combining pixel shifting algorithm with the depth-data, the input end is a AV signal and output end is a VGA signal. The processed video stream are outputting to a HMD (head mounted display-CybermanTM) with two VGA inputs, so each eye receives a different image from the FPGA platform.

Result / Conclusion：
The result from Matlab verified our assumption about small light angle reflecting intensity can be used as a factor for acquiring depth information from endoscopic images. So we can use this data to shift each pixel according to its light reflecting intensity, and then generating stereo images.

論文電子檔著作權授權書 ...............................
論文審定同意書 .......................................
中文摘要...........................I
Abstract......................... II
目錄............................ III
圖目錄.............................V
第一章 緒論 ...................... 1
1.1什麼是內視鏡 .................. 1
1.1.1內視鏡的原理 ................ 1
1.1.2內視的架構 .................. 2
1.1.3內視鏡的發展與應用 .......... 3
1.2立體視覺原理 .................. 5
1.2.1腦海中的視覺成像 ............ 7
1.2.2立體視覺的發展 .............. 8
1.2.3立體視覺應用 ............... 11
1.3深度資訊理論 ................. 15
1.4現場可程控邏輯閘陣列 ......... 16
1.4.1FPGA相關知識 ............... 16
1.4.2實現數位系統之方式 ......... 16
1.4.3 FPGA規劃方式 .............. 19
1.4.4 FPGA設計流程 .............. 20
1.5研究動機/目的 ................ 22
第二章 實驗設計與流程 ........... 24
2.1設計考量: .................... 24
2.2設計重心: .................... 25
2.3設計流程 ..................... 27
2.4實現步驟 ......................28
第三章演算法 .....................30
3.1深度資訊演算法 ............... 30
3.1.1研究假設 ................... 30
3.1.2 Matlab 模擬 ............... 31
3.2系統架構: .................... 33
第四章 結果與討論 ............... 37
4.1硬體架構 ......................37
4.2 Matlab模擬結果 .............. 39
第五章 討論 ..................... 42
第六章 結論 ..................... 44
參考文獻......................... 45

[1]http://www.cupertinovet.com/endoscopy.html
[2] David F. McAllister,“ Display Technology: Stereo & 3D Display Technologies”, Department of Computer Science North Carolina State University, pp.5- 47, 2006
[3]劉榮政,“平面螢幕之立體影像設計”,國立中央大學,光電科學研究所碩士論文,2001
[4]http://axon.physik.uni-bremen.de/research/stereo/Cyclops/index.html
[5]Berlucchi, Giovanni. "Revisiting the 1981 Nobel Prize to Roger Sperry, David Hubel, and Torsten Wiesel on the occasion of the centennial of the Prize to Golgi and Cajal.". Journal of the history of the neurosciences 15 (4): pp. 369–75，2006. doi:10.1080/09647040600639013. PMID 16997764
[6]黃正宇,蔡元綱, ”由單張影像產生立體視覺”,大同大學,專題報告,2009
[7]劉榮政,“平面螢幕之立體影像設計”,國立中央大學,光電科學研究所碩士論文,2001
[8]http://cn.made-in-china.com/showroom/shyxsg/product-detailMJnQakfGDmvP/3D%E7%AB%8B%E4%BD%93%E7%9C%BC%E9%95%9C%EF%BC%88YX-1508%EF%BC%89.html,上海影利光学材料有限公司
[9]Jeroen Smit, Richard Kleihorst, Anteneh Abbo, Jan Meuleman1 and Gerard van Willigenburg. “Real time depth mapping performed on an autonomous stereo vision module”, Wageningen, The Netherlands，pp306-310，2004
[10]Liang Zhang, Wa James Tam , “ Stereoscopic image generation based on depth images for 3D TV”, IEEE Transactions on Broadcasting, 51(2), pp.2-8, 2005
[11] Y. Ruichek, “A hierarchical neural stereo matching approach for real-time obstacle detection”, in IEEE Intelligent Transportation Systems Proceedings, vol.1 pp299 - 304，October 2003.
[12] M. Hariti, Y. Ruichek, and A. Koukam, “A voting stereo matching method for real-time obstacle detection”, in IEEE Robotics and Automation Proceedings, ICRA, vol. 2 pp1700-1704, September 2003.
[13] S. Kagami, K. Okada, M. Inaba, and H. Inoue, “Design and implementation of on body real-time depth map generation system”, in IEEE Robotics and Automation Proceedings, ICRA, vol. 2 pp1441-1446, April 2003.
[14] Y. L. Murphey, J. Chen, J. Crossman, J. Zhang, P. Richardson, and L. Sieh,，”DepthFinder, a real-time depth detection system for aided driving”, in IEEE Intelligent Vehicles Symposium Proceedings,pp122-127，October 2000.
[15] S. K. Park and I. S. Kweon, “Robust and direct estimation of 3-d motion and scene depth from stereo image sequences”, Pattern Recognition, vol. 32(9), pp.1713-1728, 2001.
[16] M. I. Fanany and I. Kumazawa, “A neural network for recovering 3d shape from erroneous and few depth maps of shaded images”, Pattern Recognition Letters, vol. 25(4), pp. 377-389, 2004.
[17] Canesta, “Canesta infrared sensor chips”, http://www.canesta.com, 2004.
[18] A. B. Robert Lange, Peter Seitz and S. Lauxterman, “Demodulation pixels in ccd and cmos technologies for time-of-fight ranging”, in IEEE Virtual Reality roceedingsIST /SPIE International Symposium on Electronic Imaging, pp177-188，January 2000.
[19]Yen–Liung Lai, “FPGA Implementation of Stereo Vision and Integration Hardware Design of a Low Vision Aid”, National Yang – Ming University, Institute of Biomedical Engineering, August 2008
[20]Ming-Bo Lin, “Digital System Designs and Practices Using Verilog HDL and FPGAs”, John Wiley & Son, 2008.
[21]Altera Corporation, “Cyclone III Device Handbook”, Version 1, March 2007.
[22]Altera Corporation, “Cyclone II Device Family Data Sheet”, Version 2.0, July 2005.
[23]R.O. Amos and C .R. Jagath, “FPGA Implementations of Neural Networks”, Springer, 2006.
[24]K. Steve, “Advanced FPGA Design Architecture Implementation and Optimization”, John Wiley & Son, 2007.
[25]http://www.olympus-global.com/en/news/2004b/nr0411263dsyse.html?&print, “Olympus News Release:Newly Developed "High-resolution 3D System" will Ease Eyestrain among Surgeons”
[26]ALTERA DE2-70 User Manual
[27]Ji-Min Li,“Design and Implementation of Real-time Image Processing of the Low Vision Aid Based on FPGA”, National Yang-Ming University, Institute of Biomedical Engineering，Master Thesis, 2008