利用顯著區域檢測與可變形模型從事鼠腦中風梗塞影像之研究

Using Salient Region Detection and Deformable Models to Analyze Infarct Regions in Rat Brain Images

林瑋祐(Wei-You Lin)

指導教授 ： 張恆華

國立臺灣大學/工學院/工程科學及海洋工程學研究所/碩士(2019年)

https://doi.org/10.6342/NTU201902948

摘要

近年來的研究顯示，中風為造成國人死亡的主要原因之一。而腦中風為腦血管阻塞或自發性破裂而造成腦部功能失常的疾病，分為缺血性及出血性，造成腦部缺乏養分及氧氣而導致腦損傷或死亡。因此在臨床研究中，腦中風為關注的議題，在臨床實驗模型中，大多使用齧齒動物之實驗影像作為研究依據。醫生須手動圈選梗塞部分，相當耗時，因此本研究提出一套演算法自動分割鼠腦梗塞區域。
本研究主要以缺血性中風鼠腦為研究對象，針對大腦動脈阻塞的大鼠為實驗對象，並以經2，3，5－氯化三苯基四氮唑染色的原始大腦影像為研究影像。本演算法首先透過顯著區域之偵測進行背景去除，之後進行伽瑪校正，得到乾淨且明亮的鼠腦影像；再進行大鼠腦部中線偵測，並且透過CIELab顏色辨別加入可變形模型進行中風區域分割。
本研究使用47隻大鼠，總共切割出348張大鼠腦部中風影像。實驗結果顯示，本研究之方法能有效的切割出鼠腦中風區域，Dice值84.8， Jaccard值為75.9， Conformity值為69.6比其他方法優秀，與醫師所切割出之結果相近。

關鍵字

缺血性中風 ； 2，3，5－氯化三苯基四氮唑 ； 顯著區域偵測 ； 伽瑪校正 ； 中線偵測 ； 中風區域分割

並列摘要

According to recent research, stroke is one of the causes of death nationwide. Stroke is a disease of brain dysfunction caused by cerebral vascular occlusion or spontaneous rupture, which can be divided into ischemic and hemorrhagic. Both types of stroke lead to nutrient and oxygen deprivation in the brain, which ultimately cause brain damage or death. Therefore, stroke is a topic of clinical concern. Usually different types of imaging modalities of rodents are used as the research basis. It is quite time consuming for researchers to manually circle the infarct. This thesis proposes an algorithm to automatically segment the rat cerebral infarction area.
This study focuses on ischemic stroke by using rat with cerebral artery occlusion as experimental subjects and examines the images of their brains obtained from staining with 2,3,5-triphenyl tetrazolium chloride (TTC). Firstly, the algorithm removes image background via the salient region detection method. After removing the background, we perform gamma correction and obtain the clean and bright brain image, after that, we can do the brain center line detection, and the segmentation of the stroke region is based on a deformable model via a CIELab distinguishing strategy.
In this study, 47 rats were used to produce a total of 348 rat brain stroke images. The experimental results showed that this study can effectively segment the rat brain stroke images, and the Dice = 84.8, Jaccard = 75.9, Conformity = 69.6 is better than other methods and closer to the estimates by doctors.

並列關鍵字

ischemic stroke ； 2,3,5-triphenyl tetrazolium chloride (TTC) ； salient region detection ； gamma correction ； center line detection ； stroke region segmentation

參考文獻

[1] The top 10 causes of death. [Online]. Available: https://www.who.int/topics/cerebrovascular_accident/en/

[2] J. Bamford, P. Sandercock, L. Jones, and C. Warlow, "The natural history of lacunar infarction: the Oxfordshire Community Stroke Project," Stroke, vol. 18, no. 3, pp. 545-551, 1987.

[3] J. F. Schenck, "The role of magnetic susceptibility in magnetic resonance imaging: MRI magnetic compatibility of the first and second kinds," Medical physics, vol. 23, no. 6, pp. 815-850, 1996.

[4] J. Kandampully and D. Suhartanto, "Customer loyalty in the hotel industry: the role of customer satisfaction and image," International journal of contemporary hospitality management, vol. 12, no. 6, pp. 346-351, 2000.

[5] J. B. Bederson, L. H. Pitts, S. Germano, M. Nishimura, R. Davis, and H. Bartkowski, "Evaluation of 2, 3, 5-triphenyltetrazolium chloride as a stain for detection and quantification of experimental cerebral infarction in rats," Stroke, vol. 17, no. 6, pp. 1304-1308, 1986.