隨著電腦斷層掃描與核磁共振成像的發展，醫師可以經由人體與器官的二維切片影像做為診斷依據，這些醫學成像的臨床應用在醫學診斷與治療技術上有很大的進步，然而二維影像僅能提供部份的解剖訊息，醫師很難透過堆疊二維影像來建構三維空間結構。
　　本研究利用Python與WebGL開發出基於網路的互動式三維立體重建系統，藉由取得使用者上傳的DICOM檔案，系統可將其轉換成可視化影像，並將可視化影像重建為三維立體模型。本系統的三維重建過程非常快速，十秒內即可將900張512*512的CT影像重建完成，為了讓幀率（FPS）保持在60赫茲，使用者也能在影像精緻度與繪製速度之間做取捨，因此本系統在任何硬體設備上擁有一定的流暢性，不僅如此，所有的系統功能被設計為即時控制，使用者能控制系統的互動式介面來旋轉、平移或者縮放該模型，並經由系統的參數調整功能，將三維模型以不同的型式呈現，以便於醫師分析該切片影像的特徵。
　　本研究結合了雲端運算服務，因此使用者可經由電腦、手機以及平板等裝置連接網路進入系統平台，並配合本系統的功能設計，給予醫師一個方便分析三維模型的使用工具，進而被應用於手術模擬、醫學教學以及定位放射治療上，在未來有著高度的發展性。

　　With the development of CT and MRI, physicians can use two-dimensional slices of human body and organs as the diagnosis basis. The clinical application of these medical imaging has made a great progress in medical diagnosis and treatment technology. However, two-dimensional tomographic images can only express anatomical information of a certain section. It is difficult for physicians to establish a three-dimensional spatial structure through only two-dimensional tomography images.
　　The proposed research develops a web-based interactive three-dimensional reconstruction system by Python and WebGL. With the obtainment of the DICOM files uploaded by user, the proposed system can convert it into a visual images and can reconstruct the visual images into a three-dimensional model. The reconstruction of the proposed system is rapid, and it can complete the reconstruction of 900 CT images with a size of 512*512 within 10 seconds. To keep the frame rate at 60 Hz, the users can choose between image refinement and drawing speed. Thus, the proposed system can certain fluency on any hardware device. Moreover, all the system functions are real-time. Users can rotate, translate or scale the model with the interactive interface of the system. The three-dimensional model can be displayed in different form by setting the parameter so that physicians can analyze the feature of the slice images.
　　The proposed system integrates cloud computing and enables users to analyze 3D data with portable devices via Internet. With the proposed system, medical experts could analyze three-dimensional anatomy easily and conveniently any time anywhere. The system can be applied to surgical simulation, medical teaching, and positioning radiation treatment, and the proposed system has great potential in clinical usages in the future.

[1] 政策研究指標資料庫PRIDE. https://pride.stpi.narl.org.tw/index/v3/
indcountry/15.
[2] 何謂「深度圖」(Depth Map)? http://www.i-art3d.com/TC/About_Depth.
htm.
[3] Christian Sigg and Markus Hadwiger. Fast Third-Order Texture Filtering. In GPU
Gems 2: Programming Techniques for High-Performance Graphics and General-
Purpose Computation, chapter 20, pages 313–329. Addison-Wesley, 2005.
[4] Philippe Thévenaz, Thierry Blu, and Michael Unser. Interpolation Revisited. IEEE
Transactions on Medical Imaging, 19(7):739–758, 2000.
[5] Ravish Mehra and Dr. Subodh Kumar. Real-Time Rendering Of Mesostructures.
page 8, 2014.
[6] http://blog.xuite.net/kerma2100/wretch/117019408-9901Q77Q79%EF%
BC%8C9901Q65Q66%E8%A9%B3%E8%A7%A3.
[7] Daniel Ruijters and Philippe Thévenaz. GPU Prefilter for Accurate Cubic B-Spline
Interpolation. The Computer Journal, 55(1):15–22, 2012.
[8] 吳其勳. 徹底了解Cloud Computing 部署模式. https://www.ithome.com.tw/
article/93013.
[9] Tiecheng Zhao, Robert James Taylor, Gang Li, Junnan Wu, and Chunguang Jia.
Cloud-based medical image processing system with access control. https://
patents.google.com/patent/US8682049, Feburary 2012.
[10] 張智凱. DICOM 影像格式標籤(Tags) 分析. http://imediface.blogspot.
tw/2012/01/dicom_15.html, 2012.
[11] https://goo.gl/images/gL2M1V.