I have got cerebral images from Digital Subtraction Angiography.
Blood vessels should be segmented.
I've applied Contrast-limited adaptive histogram equalization and Gaussian filter to the attached image. figure shows image before and after processing with their histogram.
What kind of pre-processing\ processing could taken place Using MATLAB to the image to be prepared for segmentation\ thresholding?
Any suggestions about a segmentation method?
I've provided a link to the original image.
References and links to similar subjects are very helpful.
https://copy.com/NcKyPHE9UD4k
try these two methods.
Truc, P.H., Khan, M., Lee, Y.K., Lee,S., Kim, T.S., “Vessel enhancement filter
using directional filter bank”, Computer Vision and Image Understanding,
pp.101–112, 2009.
Frangi, A., Niessen, W., Vincken, K., Viergever, M., “Multiscale vessel
enhancement filtering,” in: Medical Image Computing Computer-Assisted
Intervention, LNCS, vol. 1496, pp. 130–137, 1998.
Directional filter bank are very useful to enhance vessel picture.
Try with different filters, different convolution window sizes and variable number of itterations. You have to find the most appropriate parameters for these type of images. You need some experimentation. You can also download our filtering toollbox from http://www.medinfo.cs.ucy.ac.cy/. Goto the downloads tab under toolboxes. We have a number of filtering methods incorporated where you can adust the above parameters.
The image you are showing in your example has a better histogram after filtering but it looks like the edges are smoothed too much. You donot want the edges too be smoothed much eithewise the segmenattion algortihm will not work correctly. For the segmentation try snakes or level sets. Se also our papers on segmentation.
@Hossein,
@Christos ,
Do you know any available material, references about image preprocessing\processing using MATLAB ?
You may download the Despeckle filtering software toolbox from http://www.medinfo.cs.ucy.ac.cy/, under downloads tab, toolboxes.
This is an executable software with a lot of filtering methods, pre-processing, texture analysis and image quality evaluation. There is also a different toolbox with a lot of matlab functions available. Please also see following publications:
1) C.P. Loizou, C. Theofanous, M. Pantziaris, and T. Kasparis, “Despeckle filtering software toolbox for ultrasound imaging of the common carotid artery”, Computer Methods and Programs in Biomedicine, no. 114, pp. 109-124, January 2014
2) C.P. Loizou, M. Pantziaris, C.S. Pattichis, and I. Seimenis, “Brain MR Image normalization in texture analysis of multiple sclerosis”, J. Biomed. Graph. & Comput., vol. 3, no.1, pp. 20-34, 2013.
3) C.P. Loizou, C.S. Pattichis, M. Pantzaris, T. Tyllis, and A.N. Nicolaides, "Quality Evaluation of Ultrasound Imaging in the Carotid Artery Based on Normalization and Speckle Reduction Filtering", Medical and Biological Engineering and Computing, vol. 44, issue 8, pp. 405-413, 2006.
4) C.P. Loizou, C.S. Pattichis, C.I. Christodoulou, R.S.H. Istepanian, M. Pantzaris, and A.N. Nicolaides, "Comparative Evaluation of Despeckle Filtering in Ultrasound Imaging of the Carotid Artery", IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 52, issue 10, pp. 1653-1669, 2005.
5) Despeckle Filtering Algorithms and Software for Ultrasound Imaging", C.P. Loizou, and C.S. Pattichis, Morgan & Claypool Publishers, CA, USA, 2008. (Book)
I am doing the same study for Coronary Angiograms. As the preparation technique I ve applied the homomorphic filter to reduce the non-uniform illumination, after that I normalized the image by stretching it mean and variance to a desiered level. you can download my reseach papers from this site. have a look at them. after applying this filter you can get ehanced uniformly illuminated images. I also have same results. Now try to do segmentation. I got an adviced from my supervisor to look at Active Contour Models.
Hope this will help you
Hello Everyone,
I am from the remote sening community with interest in applying some of the tools and algorithms for image enhancements and processing tn the medical imaging field. Because my background is not in the medical field and I am not up to date on the state of the art I might suggest things that are obivous to you, if so, sorry.
For this application I would suggest using local, intermediate, and regional size high pass spatial filters after the image data have been contrast enhanced as shown above. These images can be either used as input to a classification software package and/or used to threshold the image. The high pass filters should remove the lower frequencies and enhance the higher frequencies depending on the size of the filters which means that veins will be the focus of the output results.
I hope I am not to far out in 'left field' and that some of this makes sense.
Pat
- Badges
- Science topic