目的：本研究的目的是要研發出一個精準的三維影像切割繪圖方法，來清楚地強化影像的模糊邊界，並引用切割的成果在強度調控放射治療的先期腫瘤定位處理，如此可以完成醫師經常所提的確保強度調控放射治療腫瘤劑量提升的特性之要求目標。材料和方法：由於小波技術具有多重尺度與多重解析的特性，使小波方法毫無疑問地成為眾所皆知最好的梯度切割方法之一，經由延伸標準離散小波轉換到二維空間，而藉此標準離散小波轉換演算法推展出一個新的邊界檢測方法，因為影像處理的實際需要，我們使用二維離散週期小波轉換來代替傳統的二維小波函數來產生二維的邊界檢測濾波器，最後套用杜比其小波轉換導出小波遮罩，來完成二維影像處理的迴旋捲積運算，如此一來使我們得以依不同尺度的邊綠檢測濾波器來做影像切割。線性插補法則與邊界連接與邊緣細線化法則亦被使用來完成三維的描繪組像。文中使用一張設計流程圖來說明我們所提出的三維治療目標定位的影像重組描繪方法之流程，並且使用不同的醫學影像來驗證我們所提方法的妥當性並將結果與傳統的區域成長切割法作比較，以呈現我們所提的方法不僅能達到理想的準確度並發現有一些對未來醫學有趣且實用的運用。結果：在驗證我們方法的影像資料當中有兩組電腦斷層掃描攝影像與一組核磁共振影像，其中第一組是人體胸腔影像資料，我們使用它來分離右肺，當作特徵的萃取，在圖三、四、五中呈現出小波切割結果，同時將區域成長切割結果描繪出來作為比較，而第二組實驗是一個腦中患有腦下垂體腫瘤的女性病患的電腦斷層掃描攝影資料，結果呈現腫瘤的大小形狀，在腦中的方位與它跟頭部的相關位置，我們都可清楚查看，圖六、七中清楚地呈現我們切割的成果，具有提供腫瘤明確尺寸與位置的能力，而第三組實驗是核磁共振影像，我們呈現如何使用著色技巧來強調重要部位，在圖十(d)中的三維重組影像顯示使用不同顏色來強調腦部與頭臚之間的相關位置。在所有實驗當中不僅達到我們想要的準確度並且獲得加入透明效果可以旋轉不同角度與使用不同顏色來強調重要部位的三維切割影像，結果可讓醫師更精準且有效率的從事手術或治療。原本有些需靠醫師自行想像組合的三維影像經由此一方法切割組像後便可清楚識別，圖像特徵清楚分辨，位置精確定義，相關位置關係可以清楚明白。這些技術在放射治療上都是相當重要的。結論：我們所提出的三維治療目標定位切割演算法可精準有效地協助醫師作傳統治療或強度調控放射治療的導引，我們發現三維小波切割描繪法不僅達到我們想要的準確性並且會在臨床醫學影像處理中具有相當地用途，因此我們相信所提的演算法將會在未來的影像導引放射治療(IGRT)與其他的醫療應用上扮演相當重要的角色。

Purpose: This study aimed to develop an advance precision three-dimensional (3-D) image segmentation algorithm to enhance the blurred edges clearly and then introduce the result onto the intensity modulated radiotherapy (IMRT) for tumor targeting. This will achieve what physicians usually demand that tumor doses escalation characteristics of IMRT. Materials and Methods: No doubt the Wavelet method is well known to be one of the best gradient segmentation methods due to its multi-scale and multi-resolution capabilities. We extend the standard Discrete Wavelet Transform (DWT) into two dimensions. By applying this 2-D DWT algorithm one may develop a new edge detection method. But for practical purposes, we replace the traditional 2-D wavelet functions with a 2-D discrete periodic wavelet transform (2-D DPWT). Finally, by using Daubechies wavelet transform function to form the wavelet masks for convolution operation in image processing, it allows us to detect edges at different scales for segmentation. Linear interpolation and edges linking and thinning algorithm were then used to form 3-D renderings. A proposed algorithm flowchart designed for this precision 3-D treatment targeting was introduced in this paper. Different medical images were used to test the validity of our method. We then compared the results with a traditional region growing segmentation approach to show that our 3-D renderings not only achieved the precision we sought but also has many interesting applications that shall be most useful to the medical practice. Results: Two CT scan data and one MRI data of different medical images were used to verify the validity of our proposed method. One CT scan dataset from a human chest was test first to extract the right lung for feature extraction. Its wavelet segmentation result was shown in the Figs. 3, 4, 5 and the region growing method was also shown for comparison. The second dataset was a CT scan of a female patient with a pituitary tumor in her brain. The size and shape of the tumor, its orientation with the brain, and the position relative to the head are all now clearly seen. Figures 6, 7 clearly demonstrated the results of our segmentation capable of providing an outstanding positioning of the tumor. In the third experiment of an MRI data, we demonstrate how colors were used in our 3-D renderings to enhance the important parts. 3-D reconstruction of the brain in a partial skull is shown in Fig. 10 (d) with different colors to emphasize their relative positions. On all examples of medical images we created, not only desired precision had been achieved, we are also able to create rotation of the objects with transparent effects to obtain its 3-D images of different angles. They will allow physicians to conduct surgery or treatment much more accurately and effectively. Many 3-D images physicians usually only compose by imaginations, now all visible under our 3-D processing. Features are now clearly identified, locations pinned down exactly, and relative orientations are now well understood. These are all vital for radiation therapies. Conclusion: The 3-D targeting segmentation we proposed allows physicians to improve the traditional treatments or IMRT much more accurately and effectively. Our precision 3- D wavelet segmentation rendering shall be very useful in practice for medical images. They will play an important role in image-guided radiotherapy (IGRT) and many other medical applications in the future.