ut ut

Laboratory for Image & Video Engineering

M-FISH Chromosome Imaging Database

The M-FISH Chromosome Imaging Database was developed by former LIVE PhD student Dr. Wade Schwartzkopf under the guidance of Dr. Kenneth Castleman at Advanced Digital Imaging Research, LLC, Friendswood, TX. We acknowledge the NIH funding and IRIS International for their support in this endeavor.

We have decided to make the M-FISH Chromosome Imaging Database available to qualified researchers in the community free of charge. To obtain a copy of this database, please contact former LIVE PhD student Dr. Mehul Sampat (mehul.sampat@ieee.org). Kindly indicate your university/industry affiliation and a brief description of how you plan to use the database. Please note that this database cannot be used for commercial purposes.

If you use this database in your research, we kindly ask that you reference the following articles:

[1] M. P. Sampat, A. C. Bovik, J. K. Aggarwal, and K. R. Castleman, "Supervised parameteric and non-parametric classification of chromosome images", Pattern Recognition, Vol: 38 No: 8, August 2005, pages: 1209-1223
[2] W.C. Schwartzkopf, A.C. Bovik and B.L. Evans, "Maximum likelihood techniques for joint segmentation-classification of multi-spectral chromosome images", IEEE Transactions on Medical Imaging, Vol: 14 No: 12, December 2005, pages: 1593-1610
[3] H. Choi, A.C. Bovik and K.R. Castleman, "Feature normalization via expectation maximization and unsupervised nonparametric classification for M-FISH chromosome images" IEEE Transactions on Medical Imaging, Vol: 27 No: 8, August 2008, pages: 1107-1119


Images of chromosomes may be obtained using a number of methods. One such method, is Multiplex Fluorescence In Situ Hybridization (M-FISH) [1,2] which is a chromosome imaging technique developed in the mid 1990's. The first paper on the M-FISH technique was published in 1996 by Speicher et al. [1] and it revolutionized chromosome imaging. In this technique chromosomes are labeled with 5 fluors (dyes) and a fluorescent DNA stain called DAPI (4 ,6-Diamidino-2-phenylindole). DAPI attaches to DNA and thus labels all chromosomes. The fluors attach to specific sequences of DNA. With MFISH a unique combination of fluors is assigned to each chromosome type. That is, each class of chromosomes absorbs a different combination of fluors [1]. Thus M-FISH is based on a combinatorial labelling strategy. This strategy provides an easy way to label chromosomes in a multiplex fashion, as each fluor is either present(1) or absent(0) [1,3]. Also, at least five distinguishable fluors are needed for combinatorial labelling to uniquely identify all 24 chromosome types as the number of useful combinations of N fluors is 2*N - 1 [1,3]. The central idea in M-FISH is that each chromosome is labeled by a unique combination of the five fluors.

[1] M. Speicher, S. Ballard, D. Ward, Karyotyping human chromosomes by combinatorial Multi-Fluor FISH, Nat. Genetics 12 (1996) 368--375.
[2] M. Beau, One FISH, two FISH, redFISH, blue FISH, Nature Genetics 12 (1996) 341--344
[3] M. Speicher, S. Ballard, D. Ward, Computer image analysis of combinatorial Multi-Fluor FISH, Bioimaging 4 (1996) 52--64.

The M-FISH image database web pages are being maintained by Mehul Sampat