Unit of Medical Technology and Intelligent Information Systems

Sitemap



	Modelling of biological tissues and systems
	Automated Diagnosis
	Bioinformatics
	Patient Monitoring Systems
	Biomagnetism

[Back]

Chromosome Image Processing

Chromosome analysis is an essential procedure for detecting genetic abnormalities. Traditionally, cells are classified according to their karyotype, which is a tabular array in which the chromosomes are alligned in pairs. Chromosome images are analyzed by experts to obtain vital information about the health of an individual. Normal cells have 22 pairs of chromosomes and 2 sex determinative (XX: women, XY: men). Abnormal cells may have an excess or a deficit of chromosomes and/or structural defects which depict an exchange of genetic material.

However, manual examination of these images is a laborious and difficult task requiring skilled lab technicians. Many attempts have been made to automate parts of the chromosome image analysis procedure. Our main focus is to establish image processing methods that will provide automated karyotyping making the decision of the expert more objective. More specific, the method that we have already developed can be separated into two main stages:

Segmentation

This stage aims to define regions (groups of pixels) that share same characteristics such as color. We have used the watershed transform which is a popular segmentation method coming from the field of mathematical morphology.
The description of this transform is quite simple: we must consider the image as a topographical relief, where the height of each point is related to its grey level. Then imagine rain falling on the terrain, the watersheds are the lines that separate the catchment basins that form.
An example of the application of the Watershed Transform to a chromosome image is shown below.

Classification

In this stage we must classify each segmented region into 1-24 categories (1-22, X, Y). We use a region classification method based on the Bayes rule. The result of this stage is a classification map where each segmented region is classified to 1-24 chromosome classes. A map of this type is shown below:

People: Petros Karvelis

Useful Links
	Petros Karvelis http://www.cs.uoi.gr/~pkarvel
Articles in journals:
[1].	P. S. Karvelis, A. Tzallas, D. I. Fotiadis, and I. Georgiou, “A Multichannel Watershed-Based Segmentation Method for Multispectral Chromosome Classification”, IEEE Transactions on Medical Imaging, vol. 27, no. 5, pp. 697-708, 2008.
[2].	P. S. Karvelis , D. Fotiadis, D. Tsalikakis and I. Georgiou, “Enhancement of Multichannel Chromosome Classification Using a Region Based Classifier and Vector Median Filtering,” IEEE Transactions on Information Technology in Biomedicine, vol. 13, νο. 4, pp. 561 – 570, Apr 2009.
Articles in conferences:
[1].	P. S. Karvelis, and D. I. Fotiadis, "A region based decorrelation stretching method: application to multispectral chromosome image classification ", 2008 IEEE International Conference on Image Processing, (ICIP 2008), San Diego, California.
[2].	P. S. Karvelis, and D. I. Fotiadis, "Enhancement of Multispectral Chromosome Image Classification Using Vector Median Filtering", 15th International Conference on Conceptual Structures (ICCS 2007), Sheffield, England, 2007.
[3].	P.S. Karvelis, D.I. Fotiadis, A.T. Tzallas, and I. Georgiou, “Region Based Segmentation and Classification of Multispectral Chromosome Images,” 20th IEEE International Symposium on Computer-Based Medical Systems, (CBMS 2006), June 20-22, Maribor, pp. 251-256, 2007
[4].	P. S. Karvelis, D. I. Fotiadis, Analysis of Cytogenetic Abnormalities using Multispectral Chromosome Images, International Special Topic on Information Technology in Biomedicine, (ITAB 2006), Oct 26-28, 2006.
[5].	P. S. Karvelis, D. I. Fotiadis, I. Georgiou, and M. Syrrou, A Watershed Based Segmentation Method for Multispectral Chromosome Images Classification, Proceedings of the 28th IEEE EMBS Annual International Conference, (EMBC 2006), pp. 3009-3012, New York City, USA, Aug 30-Sept 3, 2006
[6].	P. S. Karvelis, D. I. Fotiadis, M. Syrrou, I. Georgiou, Segmentation of chromosomes images based on a recursive watershed transform, Proceedings of the 3rd European Medical & Biological Engineering Conference, (EMBEC 2005), pp. 497-502, Prague, 20-25 November 2005.

[Back]