北京生物醫(yī)學(xué)工程

基于距離變換與路徑規(guī)劃的骨架提取算法

Extraction of skeleton based on distance transform and path planning

作者：張超蘆勤羅述謙

單位：首都醫(yī)科大學(xué)生物醫(yī)學(xué)工程學(xué)院(北京100069)

分類號(hào)：

出版年·卷·期（頁碼）：2012·31·6（551-555）

摘要：

目的骨架具有與原始物體相同的拓?fù)渑c形狀信息，能夠有效地描述物體，在醫(yī)學(xué)圖像處理中有很好的應(yīng)用前景，但傳統(tǒng)基于距離變換的骨架提取難以保證骨架的連續(xù)性，因此引入骨架候選點(diǎn)概念以解決連續(xù)性問題。本文提出一種基于距離變換和路徑規(guī)劃的骨架提取算法。方法首先利用距離變換后的約束條件（局部距離變換最大值和局部距離變換梯度的模的最小值），選擇骨架候選點(diǎn)，利用動(dòng)態(tài)規(guī)劃中最短路徑原理連接骨架候選點(diǎn)。結(jié)果本文算法產(chǎn)生的結(jié)果與標(biāo)準(zhǔn)骨架對(duì)比，平均相似度達(dá)94%以上。結(jié)論基于距離變換和路徑規(guī)劃的骨架提取算法很好地保護(hù)了骨架的拓?fù)湫院瓦B續(xù)性，無需過多的后處理，并引入了相似度的概念來評(píng)價(jià)骨架算法。

Objective The topology and shape information of the skeleton are similar with the original object.The skeleton can effectively describe the object and has great application prospect in medical image processing.Traditional skeletonization based on distance transform cannot guarantee the connectivity of the skeleton,so we introduce candidate skeleton points (CSP) to solve the connectivity problem.In this paper,we propose a well-connected skeleton algorithm based on distance transform and path planning.Methods According to the constraints of distance transform (local maximum value of distance transform and local minimum points of the norm of distance transform gradient),candidate skeleton points are located,and the shortest path between CSPs can be determined using path planning theory.Results The comparison between the results of our proposed method and the standard centerline of the test images shows that the average of similarity measurement (SM) is above 94%.Conclusions The proposed algorithm based on distance transform and path planning can generate a connected skeleton with well preserved original object’s topology without requiring a post-processing step,and SM is defined to evaluate the accuracy of the skeleton.

參考文獻(xiàn)：

［1］Blum H. A transformation for extracting new descriptors of shape［M］//Wathen-Dunn W,eds. Models for the Perception of Speech and Visual Form. Cambridge，MA:MIT Press,1967：362-380.
［2］Xu Y,Zhang H,Li H,et al. An improved algorithm for vessel centerline tracking in coronary angiograms［J］. Comput Meth Prog Bio，2007,88:131-143.
［3］廖志武. 2-D骨架提取算法研究進(jìn)展［J］. 四川師范大學(xué)學(xué)報(bào):自然科學(xué)版,2009，32(5):676-688.
Liao Zhiwu. A survey of 2-D skeletonization algorithm［J］. Journal of Sichuan Normal University：Natural Science,2009，32(5):676-688.
［3］Liao Z.A survey of 2-D skeletonization algorithm［J］. Journal of Sichuan Normal University:Natural Science,2009，32(5):676-688.
［4］Xia H,Tucker PG. Fast equal and biased distance fields for medial axis transform with meshing in mind［J］. Appl Math Model，2011，35(12):5804-5819.
［5］Ramanathan M,Gurumoorthy B. Interior medial axis transform computation of 3D objects bound by free-form surfaces［J］. Comput Aided Design,2010,42(12):1217-1231.
［6］Ding M,Tong R,Liao S,et al. An extension to 3D topological thinning method based on LUT for colon centerline extraction［J］. Comput Meth Prog Bio，2009，94(1):39-47.［7］Palágyi K. A 3D fully parallel surface-thinning algorithm［J］. Theor Comput Sci，2008，405:119-135.
［8］Lam L,Lee S. Thinning methodologies:a comprehensive survey ［J］. IEEE T Pattern Anal,1992,14(9):869-885.
［9］Jaishankar S,Pralhad RN. 3D off-Line path planning for aerial vehicle using distance transform technique［J］. Procedia Computer Science,2011，4:1306-1315.
［10］Couprie M,Coeurjolly D,Zrour R. Discrete bisector function and Euclidean skeleton in 2D and 3D［J］. Image and Vision Computing,2007，25:1543-1556.
［11］Boxer L. Efficient Computation of the euclidean distance transform［J］. Comput Vis Image Und,2000，80:379-383.
［12］Fu Y,Liu H,Wang S,et al. Skeleton-based active catheter navigation［J］. Int J Med Robot,2009,5(2):125-135.
［13］劉大鵬,孟曉林,張煜,等. 虛擬結(jié)腸鏡的快速中心路徑提取算法［J］. 中國醫(yī)學(xué)物理學(xué)雜志,2010，27（3）:2254-2257.
Liu Dapeng,Meng Xiaolin,Zhang Yu,et al. Fast centerline extraction algorithm in virtual colonoscopy［J］. Chinese Journal of Medical Physics,2010，27(3):2254-2257.
［14］Sabry Hassouna M,Abdel-Hakim AE,Farag AA. PDE-based robust robotic navigation［J］. Image Vision Comput，2009，27(1-2):10-18.
［15］Malandain G,Fernhdez-Vidal S. Euclidean skeletons ［J］. Image and Vision Computing，1998，16:317-327.
［16］Hassouna MS,Farag AA. Variational curve skeletons using gradient vector flow［J］. IEEE Trans Pattern Anal Mach Intell,2009,31(12):2257-2274.
［17］Bellman RE. Dynamic Programming［M］. New York:Courier Dover Publication,2003.
［18］Dijkstra EW. A note on two problems in connexion with graphs ［J］. Numer Math，1959，1(1):269-271.
［19］Rutgers University. Rutgers university tool image database［EB/OL］. ［2012-01-12］. http://www. cs. rutgers. edu/pub/sven/rutfers-tools/.
［20］蘆勤,羅述謙. 基于MR圖像的人腦尾狀核定量分析方法研究［J］. 北京生物醫(yī)學(xué)工程,2010,29(2):120-124.
Lu Qin,Luo Shuqian. Caudate nucleus volume analysis of human brain based on MR images［J］. Beijing Biomedical Engineering,2010,29(2):120-124.

服務(wù)與反饋：

【文章下載】【加入收藏】

提示：您還未登錄，請(qǐng)登錄！點(diǎn)此登錄

51黑料吃瓜在线观看,51黑料官网|51黑料捷克街头搭讪_51黑料入口最新视频