Stenosis Detection of Arteries

I have data set of CTA of heart in 2D images. I segmented the coronary arteries by FCM clustering algorithm. For detecting stenos, i implement center line algorithm and get the skeleton of arteries.Now i want to get the width of arteries. Can anyone help me in this regard that how i find width or discontinuity of arteries in MATLAB.The IMAGE OF SEGMENTED ATERIES WITH CENTRELINE IS ATTACHED HEREBY

[Fareena]

can you please share your code for fcm clustering.

[Asadullah Mumtaz]

can you please share the code this output image

Answer 1

[Jeff E]

Not sure what you mean by discontinuity, but on way to obtain an average width would be the sum of the artery mask divided by the total length of the midline skeleton.

[Image Analyst]

I second that. That should probably be good enough, particularly since the result might be a fair amount different had you just taken the radiograph from a viewpoint rotated a few degrees or at a different time in the heartbeat cycle.

[beenish Aziz]

The discontinuity means breakage in line or you may say thinness of arteries. The location where the arteries become thin indicated the stenos.

Infact i have an idea that i want to obtain a 2D area of the lumen perpendicular to centerline at each centerline point . Repeat this area computation along each of the centerline point. Plot a 1D plot with the area along the centerline. If there is a sudden drop in value in the plot, there is stenosis

But i need help in coding as i am not too good in MATLAB. I can't understand which function will help me in computing 2D area at each centerline point that is perpendicular at centerline.

Yes i have radiograph of whole heart that is rotated at different angles from viewpoint

[Image Analyst]

Brings back memories. Many years ago I got my doctorate in tracking vessels in angiograms and finding stenoses. Of course it's not easy and not just some simple chunk of a hundred lines of code that I can give you. People (like you) are still working on the problem today so that tells you it's not easy. We can't really do complicated algorithm development here because you can find such image processing algorithms here: VisionBib. Go there and see how other people have done it. But if you have smaller, more specific MATLAB questions, we can try to answer those.

Since this project will require that you get good at programming, please see this link: http://www.mathworks.com/matlabcentral/answers/8026-best-way-s-to-master-matlab

[beenish Aziz]

Thanks a lot Image Analyst to guide me properly.

[beenish Aziz]

Dear Image Analysts

i have one more question that is there any general purpose method/code that can extract our desired blob from 10 different intensity images on basis of area or intensity computation. in fact i want the code that can remove the Arota from the images at each orientation for the segmentation of coronary arteries.As i have to hard code the ranges of intensities and area which will become changed for each image. I get my ROI by using Fuzzy C-means clustering on 2D RGB Image.Please suggest me in this regard.

Answer 2

[Harsha Jawale]

please share your code