Return to search

Feature Analysis of Coronary Artery Calcification in CT Image

Detection of coronary calcification from computed tomography scans is a noninvasive examination and has great potential for heart disease diagnosis. This study proposes new features to characterize calcium lesions. Different from the traditional calcium score features, these features are generated from the texture, shape, and gray level statistics of the calcium lesions. We study the correlation between these features and acute myocardial infarction and then compare the results with traditional calcium score features.
According to the location of the lesions, patients are first divided into two groups. In one group, the myocardial infarction is located in segment of culprit lesion. For the second group, myocardial infarction is located in segment of non-culprit lesion. In comparing their means, the corresponding p-value of gray level statistics feature ¡§histogram relative smoothness¡¨ could reach 4.47E-07, which has fairly high significance.
Furthermore, in studying the differences among patients, experiment on Hypercholesterolemia provide a good result. By comparing the means of patients which are classified into Hypercholesterolemia and non-Hypercholesterolemia groups, the corresponding p-value of texture feature ¡§inverse difference moment¡¨ could reach 3.74E-04. Initially, acute myocardial infarction do not have statistically significant result However, after adding the location weighting factor for the lesions, the corresponding p-value of texture feature ¡§average of information measures of correlation¡¨ could be reduced to 8.8921E-03, and p-value of gray level statistics feature ¡§Histogram relative Smoothness¡¨ could be improved to 2.4019E-02.

Identiferoai:union.ndltd.org:NSYSU/oai:NSYSU:etd-0826109-174710
Date26 August 2009
CreatorsChang, Ta-jen
ContributorsMing-Ting Wu, Chen-Wen Yen, Pei-Chung Chen
PublisherNSYSU
Source SetsNSYSU Electronic Thesis and Dissertation Archive
LanguageCholon
Detected LanguageEnglish
Typetext
Formatapplication/pdf
Sourcehttp://etd.lib.nsysu.edu.tw/ETD-db/ETD-search/view_etd?URN=etd-0826109-174710
Rightsunrestricted, Copyright information available at source archive

Page generated in 0.009 seconds