Partial Volume Segmentation of Cerebral MRI Scans with Mixture Model Clustering

Author(s):  
Aljaž Noe ◽  
James C. Gee
Stroke ◽  
2012 ◽  
Vol 43 (suppl_1) ◽  
Author(s):  
Dean Shibata ◽  
Therese Tillin ◽  
Norman Beauchamp ◽  
John Heasman ◽  
Wadyslaw Gedroyc ◽  
...  

Introduction: Stroke mortality is doubled in people of Black African descent compared with Whites, but factors responsible for this excess are unclear. We wished to compare infarct like lesions (ILL) on MRI by ethnicity and the role of risk factors. Methods: SABRE is a UK community based multi-ethnic cohort of men and women aged 40-69 years at baseline (1988-1990), and 58-86 years at follow up (2008-2011). At follow up, a questionnaire was completed and investigations performed including resting and ambulatory BP, anthropometry, and bloods for glucose and lipids. Cerebral MRI scans were scored for infarcts independently by two readers according to the Cardiovascular Health Study protocol. Results: Of 2346 Whites, 684 attended follow up, and 590 completed cerebral MRI. Of 801 Blacks (first generation migrants of Black African descent to the UK), 232 attended clinic and 207 completed MRI. Mortality loss was greater in Whites (605, 25%) than Blacks (121, 15%)(p<0.0001), although stroke was more likely the underlying cause in Blacks (23, 19%), than Whites (43, 7%)(p<0.0001) . Baseline systolic/diastolic BP was similarly higher in Blacks than Whites in attendees (8/5 mmHg), non-responders (7/6 mm Hg), and those who died (8/5 mmHg). At follow up stroke risk factors were adverse in Blacks, apart from smoking ( table ). Prevalence of ILL was similar by ethnicity, not differing when those <65 years were analysed separately, or when those with stroke/TIA history were excluded. Associations between ILL and risk factors did not differ by ethnicity. But prescribed treatment in those with elevated clinic BP (≥140 mmHg systolic, or ≥90 mmHg diastolic) was 83% in Blacks, 63% in Whites (p<0.0001). Further, in those with an ILL, 95% of Blacks, and 69% (p<0.0001) of Whites were on treatment. Conclusion: Equivalence of ILL rates in Blacks and Whites was unanticipated, given the greater stroke mortality in Blacks. Mitigating against selective mortality as the explanation of our findings is the similar ethnic differential in baseline BP in survivors and non-survivors, the lower overall mortality in Blacks, and overall small numbers of stroke deaths. A more likely explanation is that better targeted more aggressive treatment is now occurring in Blacks than Whites, reducing their potential burden of ILL.


2013 ◽  
Vol 760-762 ◽  
pp. 1556-1561
Author(s):  
Ting Wei Du ◽  
Bo Liu

Indoor scene understanding based on the depth image data is a cutting-edge issue in the field of three-dimensional computer vision. Taking the layout characteristics of the indoor scenes and more plane features in these scenes into account, this paper presents a depth image segmentation method based on Gauss Mixture Model clustering. First, transform the Kinect depth image data into point cloud which is in the form of discrete three-dimensional point data, and denoise and down-sample the point cloud data; second, calculate the point normal of all points in the entire point cloud, then cluster the entire normal using Gaussian Mixture Model, and finally implement the entire point clouds segmentation by RANSAC algorithm. Experimental results show that the divided regions have obvious boundaries and segmentation quality is above normal, and lay a good foundation for object recognition.


2017 ◽  
Author(s):  
Nathan Olliverre ◽  
Muhammad Asad ◽  
Guang Yang ◽  
Franklyn Howe ◽  
Gregory Slabaugh

Sign in / Sign up

Export Citation Format

Share Document