本文選題：骨骼提取 + 三維分割��； 參考：《東北大學(xué)》2013年碩士論文

【摘要】：隨著我國的快速發(fā)展,國民對健康生活的渴望也日漸增強。心腦血管疾病是當(dāng)今人類的嚴重威脅之一,50歲以上中老年人更是心腦血管疾病的高發(fā)人群。我國每年有近300萬人死于心腦血管疾病,占我國每年總死亡病因的51%。因此,對心腦血管疾病的早期診斷和準(zhǔn)確診斷成為近代醫(yī)學(xué)研究的一個重要課題。計算機斷層造影(CTA),在醫(yī)學(xué)上又叫非創(chuàng)傷性血管成像技術(shù),現(xiàn)在已經(jīng)成為診斷心腦血管疾病的一種重要方法,尤其在介入治療中起著不可替代的作用。本文旨在更加精確更加快速地分離提取出頭部骨骼組織和獨立的下頜骨結(jié)構(gòu),輔助醫(yī)生進行更準(zhǔn)確的診斷。本文首先,根據(jù)人體生理特征和頭部醫(yī)學(xué)影像的特點,提出了醫(yī)學(xué)影像預(yù)處理算法和特征區(qū)域骨骼統(tǒng)計頭部分層算法。醫(yī)學(xué)影像預(yù)處理算法可以有效地去除醫(yī)學(xué)影像中的無關(guān)部分,并保留影像中有用的人體部分。特征區(qū)域骨骼統(tǒng)計頭部分層算法可以將復(fù)雜的頭部數(shù)據(jù)分成三個部分,使得針對不同部位的診斷可以僅僅只處理與之相對應(yīng)的數(shù)據(jù)集,有效地減少了數(shù)據(jù)處理量。其次,本文提出一種全新的基于區(qū)域圓形度、灰度均值和均方差的組織判別方法。提出一種全新的改進的主動輪廓模型,并將它和三維區(qū)域生長算法相結(jié)合,良好地解決了頭部骨骼組織提取問題。然后,通過大量臨床數(shù)據(jù)的實驗比較,證明了本文提出的方法優(yōu)于傳統(tǒng)的基于閾值的骨骼分割方法。再后,本文在醫(yī)學(xué)影像理論的基礎(chǔ)上提出了谷狀結(jié)構(gòu)理論,并以此為依據(jù),進一步提出了關(guān)節(jié)軟骨檢測算法,通過該算法,使得不同骨骼的分離問題得以有效地解決。最后,本文創(chuàng)新地設(shè)計了下頜骨獨立提取算法,并提出自主設(shè)計的下頜骨初始種子點快速確定算法和關(guān)節(jié)軟骨檢測算法。通過這兩個自主創(chuàng)新設(shè)計的算法,完美地解決了下頜骨提取問題,并通過大量臨床數(shù)據(jù)實驗予以驗證。
[Abstract]:With the rapid development of our country, the people's desire for healthy life is also increasing. Cardiovascular and cerebrovascular diseases are one of the serious threats to human beings today. The middle-aged and elderly people over 50 years of age are the high incidence of cardiovascular and cerebrovascular diseases. In our country, nearly 3 million people die from cardiovascular and cerebrovascular diseases each year, accounting for the 51%. of the total death cause of our country every year. Early diagnosis and accurate diagnosis of vascular diseases have become an important subject in modern medical research. Computed tomography (CTA), which is called non traumatic angiography, has now become an important method for the diagnosis of cardiovascular and cerebrovascular diseases, especially in interventional therapy. In this paper, a medical image preprocessing algorithm and a feature region skeleton algorithm are proposed for medical image preprocessing. In order to effectively remove the unrelated parts of medical images and retain the useful part of the image, the feature region skeleton statistics head stratification algorithm can divide the complex head data into three parts, so that the diagnosis of different parts can only deal with the corresponding data set and effectively reduce the amount of data processing. In this paper, a new organizational discrimination method based on regional roundness, gray mean and mean variance is proposed. A new improved active contour model is proposed and combined with three dimensional region growth algorithm to solve the problem of bone tissue extraction in the head. The method proposed in this paper is superior to the traditional threshold based bone segmentation method. After that, this paper puts forward the theory of valley structure on the basis of medical imaging theory. On the basis of this, the joint cartilage detection algorithm is further proposed. Through this algorithm, the separation problem of different bones can be effectively solved. Finally, this paper is innovative. The algorithm of mandible independent extraction is designed, and the algorithm for determining the initial seed point of the mandible is proposed and the articular cartilage detection algorithm is designed independently. Through these two independent innovative design algorithms, the problem of mandible extraction is perfectly solved and verified by a large number of clinical data experiments.

【學(xué)位授予單位】：東北大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2013
【分類號】：R816.2;TP391.41
，

本文編號：1821549