Personnel
André Khalil
University of Maine, Orono
Department of Mathematics & Statistics
324 Neville Hall
(207) 581-3911
andre.khalil@umit.maine.edu
andre.khalil@jax.org
http://germain.umemat.maine.edu/faculty/khalil/
http://www.umaine.edu/genomics/faculty/?profile=356
On the Development and Use of Rigorously Well-Defined, Quantitative, and Objective Image Analysis Tools: Keeping Up With Technological Advancements
The recent technological development of imaging and microscopy techniques has become so important (the recent arrival of the 4Pi microscope at The Jackson Laboratory (TJL) being a good example) that equally advanced, quantitative, and objective image analysis tools must be developed in order to characterize everything that this new technology has to offer. Our work is centered on the development and application of two quantitative image analysis formalisms: The Metric Space Technique (MST) and the Wavelet-Transform Modulus Maxima Method (WTMMM).
The MST is a general formalism based on the mathematical branches of Analysis and Topology, which yields an overall complexity score of the analyzed image. Instead of characterizing an image on a pixel-to-pixel basis, the MST allows one to characterize an image quantitatively, by way of bio-physically meaningful “output functions”. These output functions are user-defined and application-related, and can be as diverse as the distribution of density in the image, to the actual filamentary structure of the analyzed objects, which makes the MST a very powerful, versatile, and most importantly, adaptable image analysis tool.
The WTMMM is a wavelet-based multifractal formalism that has been developed over the past twenty years and used in signal and image analysis applications in almost all of the applied sciences, including DNA sequence analysis, finance, turbulence, astrophysics, geophysics, and bio-medical imagery. The WTMMM uses the multi-scale analyzing power of the wavelet-transform to characterize objects that have details at all scales. Image segmentation, overall morphological characterization, and modeling are examples of what can be done with the aid of the WTMMM.
Both tools are presently being developed and used for object segmentation, geometrical and morphological characterization, and (multi) fractal analysis of:
1- Chromosome territories in mouse bone marrow cell nuclei (with Kevin Mills, TJL).
2- Gene-rich vs. gene-desert parts of the piebald region on distal mouse chromosome 14 (with Lindsay Shopland, IMB).
3- Membrane surfaces (with Sam Hess, U.Maine and Josh Zimmerberg, NIH).
4- Trabecular structure and directional preferences of femoral bone in B6 vs C3H mice (with Wes Beamer and Cliff Rosen, TJL)
5- Chromosome territories in chicken bone marrow cell nuclei structure (with Steffen Dietzel, University of Munich, Germany).
Most of these image analysis projects are done in collaboration with Alain Arneodo, Ecole Normale Superieure de Lyon, France.
Recent Publications:
Khalil A, Joncas G, Nekka F. 2002. Exotic tools for the morphological analysis of HI clouds. Taylor AR, Landecker TL, Willis AG (eds). ASP Conference Proceedings, San Francisco, Astronomical Society of the Pacific. Vol. 276, p. 194-197
Khalil A, Joncas G, Nekka F. 2004. Morphological analysis of HI features I: Metric space techniques. Astrophys J 601:352-364,
Khalil A, Joncas G, Nekka F, Kestener P, Arneodo A. 2005. Morphological analysis of HI features II: Wavelet-based multifractal formalism. Astrophys J (submitted)
Joncas G, Robitaille JF, Khalil A. 2005. Morphological analysis of phase II and phase III CGPS HI features. (in prep)
Khalil A, Arneodo A, Joncas G. 2005. HI modeling in spiral and inter-arms. Astrophys J Let (in prep)
Back to Personnel
