以解剖結構形狀特徵為基礎之自動化肺區擷取及肺裂搜尋演算法

肺葉分區不僅可以做為肺病良惡性診斷的輔助資訊，也可以用來評估肺部手術後的肺功能保存狀況，或者提供醫療教育等功能。雖然醫療專業人員可藉由形狀和位置的資訊找到肺葉的邊界，然而，胸腔的全肺CT掃描通常有數百張的切片影像，以人為手動的方式進行尋找肺裂並且描繪肺葉的分界，不僅耗時而且必須耗用大量人力，因此，藉由肺葉分割自動化能有效改善人為操作時間，間接提升醫療品質。本研究演算法主要以解剖結構的形狀特徵為觀點，由肺區擷取逐步推展至肺裂搜尋。為排除非呼吸系統的器官，本研究採用波前擴張式三維區域成長演算法進行肺區擷取；接著，在左右肺分離的步驟，亦提出一種沿肺壁輪廓進行強迫分割的新演算法，此方法可有效排除左右肺交界處不屬於肺實質的組織，得到良好的分割效果；最後，考慮肺裂於空間中為板狀結構物的觀念，採用Wiemker所提出的板狀結構濾波器進行肺裂初步搜尋，再利用三維的 Neutrosophic (NS) 濾波器強化肺裂形態而得到更完整的肺裂分割。研究結果顯示，利用波前擴張式三維區域成長演算法的增長效益與肺壁輪廓進行左右肺分割的方法，我們得到良好的左右肺區擷取。基於這個良好的肺區分割，板狀結構濾波器與三維NS濾波器也順勢分割出有效的肺裂。

關鍵字

肺葉分區；三維區域成長；左右肺分離；肺裂搜尋； NS濾波器

並列摘要

Segmentation of the pulmonary lobes is important to localize parenchyma disease inside the lungs and to quantify the distribution of a parenchyma disease. Since the proposed fissure segmentation system can provide a visualization of a patient’s upper and lower lungs, it also could be incorporated in teaching software for medical professionals. Although radiologists might be able to identify lobar boundaries on CT scans, manual delineation of over hundreds CT images is unthinkable in clinical routine. Therefore, computer-aided diagnosis (CAD) is strongly desired to assist radiologists in CT image interpretations. This work proposed a fissure detection algorithm based on the physiological structure. Before the fissure detection, it is necessary to have a good lung region segmentation. Accordingly, we use 3D region growing to obtain a good lung region in the first step of the proposed algorithm. Next, we separate the lung region to right and left by following the lung wall. Finally, the fissure is segmented by using fissure filter and 3D neutrosophic (NS) filter. The experimental results show that we have proposed algorithm for fissure segmentation has good performance.

並列關鍵字

Lobe segmentation ； 3D region growing ； Right and left lung separation ； Fissure Detection ； NS filter

參考文獻

[1] Brown, M.S., McNitt-Gray, M.F., Goldin, J.G., Suh, R.D., Sayre, J.W., Aberle, D.R., 2001. “Patient-specific models for lung nodule detection and surveillance in CT images.” IEEE Transactions on Medical Imaging 20 (12), 1242–1250.

[2] Ko, J.P., Betke, M., 2001. Chest CT: “automated nodule detection and assessment of change over time – preliminary experience.” Radiology 218 (1), 267–273.

[6] Mullally, W., Betke, M., Wang, J., Ko, J., 2004. “Segmentation of nodules on chest computed tomography for growth assessment. ”Medical Physics 31 (4), 839–848.

[9] Gierga, D.P., Chen, G.T.Y., Kung, J.H., Betke, M., Lombardi, J., Willett, C.G., 2004. “Quantification of respiration-induced abdominal tumor motion and the impact on IMRT dose distributions.” International Journal on Radiation Oncology – Biology – Physics 58 (5), 1584–1595.

[11] Adler Jr., J.R., Murphy, M.J., Chang, S.D., Hancock, S.L., 1999. Image guided robotic radio surgery. Neurosurgery 44 (6), 1299–1306.

國際替代計量

以解剖結構形狀特徵為基礎之自動化肺區擷取及肺裂搜尋演算法

全文下載

主題瀏覽