Problems with Frangi's filter

4 次查看（过去 30 天）

Alessandro Russo 2016-3-14

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/273413-problems-with-frangi-s-filter

Hi all, i have a cardiac CT volume of 512x512x125, and i want to apply Frangi's filter (downloaded from file exchange http://www.mathworks.com/matlabcentral/fileexchange/24409-hessian-based-frangi-vesselness-filter) in order to extract the coronary vesselness map. The problem is, i give as input the approximate vesselness map obtained with global tresholding (a thin band of the external heart wall on which there is coronaries course, with also the ventricles in some slices, all sourrounded by black background), setting the filter parameters as found in literature for this applications (alpha=0.5, beta=10, gamma=10.000, range of sigmas variable between 0 and 2 with step of 0.05), but the result (plotted with vol3d) is the one in figure, totally incorrect.

What am i doing wrong? Of course i tried a lot of different sets of parameters, but the result is always unsatisfying. I attach the code which applies the filter, the workspace with the variables useful for compiling and the tool vol3d for visualization if wanted.

mex eig3volume.c;
options.BlackWhite=false;
options.FrangiScaleRange=[0 3];
options.FrangiScaleRatio=0.01;
options.FrangiC=10000;
options.FrangiAlpha=0.5;
options.FrangiBeta=15;
Vfiltered=FrangiFilter3D(V,options);
figure(1), vol3d('cData',Vfiltered)