blood vessel segmentation help

Hi,

I want to segment cerebral blood vessels from angiogram images?

Results which I got are very disappointing.

Any help or suggestions about image enhancement and segmentation are really appreciated.

This is an example of images to be segmented: https://copy.com/ezNK3fM6sVNs

I do some processing but it is insufficient

https://copy.com/BzpfTBTe4Foe

[Sowmiya Jagadeesan]

Hi sir I need a segmentation code sir

Answer 1

[Spandan Tiwari]

This may not be directly relevant but since you mentioned FCLAHE I wanted to point out that the ADAPTHISTEQ function in the Image Processing Toolbox is based on the CLAHE algorithm.

Answer 2

[Nicola Strisciuglio]

Here there is a full example of segmentation in retinal fundus images, which can be easily adapted to angriography images.

http://www.mathworks.com/matlabcentral/fileexchange/49172-trainable-cosfire-filters-for-vessel-delineation-with-application-to-retinal-images

[Image Analyst]

Very impressive results.

[Nisreen Sulayman]

@Nicola

Have you build a database of the results? Or do you know any matlab code to build a an indexed database according to image features??

[Nisreen Sulayman]

@ Image Analyst

Have you read my last comment

http://www.mathworks.com/matlabcentral/answers/167366-how-to-measure-circularity

Answer 3

[Image Analyst]

We don't really do complicated algorithm development here - just MATLAB programming stuff like syntax, error messages, simple program flow, etc., not something that has been the subject of countless masters and Ph.D. projects like what you're asking. Go here to see papers by those who have, and have successful results that they've published: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models

[Nisreen Sulayman]

Greeting,

I need something about intensity inhomogeneity correction especially about * using anti-diffusion model. *fuzzy contrast-limited adaptive histogram equalization (FCLAHE).

Do you have any suggestions about these two points?

[Image Analyst]

Nope - never heard of it. Sounds like some particular paper that one author wrote. There are thousands of papers and I didn't read that one. Why can't you just follow the recipe in the paper? http://matlab.wikia.com/wiki/FAQ#Can_you_program_up_the_algorithm_in_this_article_for_me_and_explain_it_to_me.3F

[Nisreen Sulayman]

I have the following result of segmentation

How can I improve the results?

1) I want to improve the appearance to get the vessels more smooth.

2) I want to extract the oval-like part of the vessels in the middle of the image.

Any suggestions?

Answer 4

[Dan]

Do you just have the single image? I would imagine that you have a stack of images comprising a volume and this additional dimension may help you....

Just my $0.02.

Answer 5

[Nisreen Sulayman]

I have the following result of segmentation

How can I improve the results?

1) I want to improve the appearance to get the vessels more smooth.

2) I want to extract the oval-like part of the vessels in the middle of the image.

Any suggestions?

Answer 6

[John]

MATLAB is a superb tool for medical image processing. It is even more powerful when coupled with other tool kits like the ITK toolkit. The kind of stuff you can do:

http://itk.org/ITKExamples/src/Nonunit/Review/SegmentBloodVesselsWithMultiScaleHessianBasedMeasure/Documentation.html

But ITK is external to MATLAB. With MATLABs external function calling capabilities you can call the right functions from precompiled ITK libraries for your platform. You could also consider the Matitk bridge software which is available at http://matitk.cs.sfu.ca/ which abstracts some parts of the interfaces with ITK

[Anand]

Or, you could look at a File Exchange entry that does the same: Hessian Based Frangi Vesselness Filter