Eilat Vardi-Gonen

Name:	Eilat Vardi-Gonen
Position:	Ph.D. student
	Computer Science Department
	The Graduate Center, CUNY
Address:	The Graduate Center, CUNY
	Computer Science Department, Room 4319
	365 5th Ave.
	New York, NY 10016
	USA
E-mail:

Research Interests:

While working with Dr. Gabor T. Herman at the Medical Image Processing Group, University of Pennsylvania, during the summer of 1998, I implemented an algorithm which uses the Metropolis Algorithm with Gibbs priors to stochastically reconstruct two-dimensional binary images from only three projection directions, see [2], [3]. In 2000, Dr. Herman, Dr. T. Yung Kong and I worked on speeding up that stochastic reconstruction algorithm for cardiac cross-section images, see [4]. In 2001, Dr. Gabor T. Herman and I worked on a stochastic segmentation algorithm using the Metropolis Algorithm with Gibbs priors. The intended application of this work was the segmentation of trabecular bone from two-dimensional magnetic resonance (M.R.) images taken in-vivo. Such images are relatively low resolution, see [5]. During that time, I did some work with Dr. Felix Wehrli to correct in-vivo M.R. images inhomogeneities, see [6].

I am currently in the Computer Science Department at The Graduate Center, CUNY. I have started working on the "real-time" interpretation of noisy signals. The application we are considering is that of improving speech signals for hearing aids. Following is a brief description of the “real-time” steps we perform in our proposed methodology. First we transform the input noisy signal to a gray image using the Short-Time Fourier Transform, STFT, a commonly used transform in many signal processing areas. Next we create a corresponding binary image using the Metropolis Algorithm with an annealing schedule where Gibbs priors are used to describe clean speech signals. We have compared different methods for estimating the prior from a training set of clean speech images. Using both the noisy gray image and the binary image, we estimate a clean gray image. One simple way of doing this is by using the binary image as a mask for the noisy gray image. Alternatively a "fuzzy" version of the binary image can be used as a mask. Finally, we estimate the clean speech signal from the clean gray image by performing the inverse signal to image transform. Preliminary work and results appeared in [7], [8] and [9].

Principal Investigator on Grants:

· Column by Column Image-Signal Interpretation, National Library of Medicine, National Institutes of Health, 1F37LM008611-01A1, 2005-2007.

Education:

Dates Attended	Degree	Department	University	City	State	Country
9/1994-5/1995			Bryn Mawr College	Bryn Mawr	PA	USA
10/1995-2/1999	B.A.	Department of Mathematics	Technion - Israel Institute of Technology	Haifa		Israel
1/2000-8/2000		Bioengineering Department	University of Pennsylvania	Philadelphia	PA	USA
9/2000-12/2000		Department of Mathematics	Temple University	Philadelphia	PA	USA
1/2001-5/2002	M.S.E.	Bioengineering Department	University of Pennsylvania	Philadelphia	PA	USA
5/2002-present	Ph.D.	Computer Science Department	The Graduate Center, CUNY	New York	NY	USA

Experience:

Dates	Position	Group	University	City	State	Country
summer 1995	Participated in "Inquiry Based Science and Math in Middle School" Program		Haverford College	Haverford	PA	USA
summer 1995	Summer Student	Laboratory of Retinal Microcircuitry	University of Pennsylvania	Philadelphia	PA	USA
8/1998-10/1998	Research Assistant	Medical Image Processing Group	University of Pennsylvania	Philadelphia	PA	USA
3/1999-7/1999	Teaching Assistant	Department of Mathematics	Technion - Israel Institute of Technology	Haifa		Israel
1/2000-8/2000	Research Assistant	Medical Image Processing Group	University of Pennsylvania	Philadelphia	PA	USA
9/2000-12/2000	Teaching Assistant	Department of Mathematics	Temple University	Philadelphia	PA	USA
1/2001-4/2002	Research Assistant	Laboratory for Structural NMR Imaging	University of Pennsylvania	Philadelphia	PA	USA
5/2002-present	Research Assistant	Discrete Imaging and Graphics Group	The Graduate Center, CUNY	New York	NY	USA

Conferences:

7th Summer School on Image Processing, Szeged, Hungary, July, 1999. (See group project here.)
Workshop on Discrete Tomography: Algorithms and Applications, Siena, Italy, October, 2000.
8th International Workshop on Combinatorial Image Analysis, Philadelphia, PA, USA, August, 2001.
Workshop on Digital Topology, New York, NY, USA, March, 2002.
2004 IEEE International Symposium on Biomedical Imaging, Arlington, VA, USA, April, 2004.
IEEE 31st Annual Northeast Bioengineering Conference, Hoboken, NJ, USA, April, 2005.
7th IEEE EMBS International Summer School on Biomedical Imaging, Berder, France, June, 2006.
IEEE Signal Processing Society - 12th Digital Signal Processing Workshop; 4th Signal Processing Education Workshop, Grand Teton National Park, WY, USA, September, 2006.

Publications:

[1] H. von Gersdorff, E. Vardi, G. Matthews and P. Sterling. Evidence that Vesicles on the Synaptic Ribbon of Retinal Bipolar Neurons Can be Rapidly Released. Neuron, volume 16(6): pages 1221-1227, 1996. Click here to see article online.

[2] B. M. Carvalho, G. T. Herman, S. Matej, C. Salzberg and E. Vardi. Binary Tomography for Triplane Cardiography. In A. Kuba, M. Samal and A. Todd-Pokropek, (eds.): Information Processing in Medical Imaging , volume 1613: pages 29-41. Springer-Verlag, Berlin, 1999.

[3] S. Matej, A. Vardi, G. T. Herman and E. Vardi. Binary Tomography Using Gibbs Priors. In G. T. Herman and A. Kuba (eds.): Discrete Tomography: Foundations, Algorithms and Applications, pages 191-212. Birkhauser, Boston, 1999.

[4] E. Vardi, G. T. Herman, and T. Y. Kong. Speeding Up Stochastic Reconstructions of Binary Images from Limited Projection Directions, Linear Algebra and its Applications: Special Issue on Discrete Tomography, volume 339: pages 75-89. Elsevier, 2001. Click here to see article online.

[5] E. Vardi and G. T. Herman. Stochastic Segmentation Using Gibbs Priors, Electronic Notes in Theoretical Computer Science, volume 46: pages 381-392. Elsevier, 2001. Click here to see article online.

[6] B. R. Gomberg, M. Fernandez-Seara, B. S. Zemel, P. K. Saha, E. Vardi, L. Loh, L. Hilaire, and F. W. Wehrli. Measurement of Trabecular Bone Volume Fraction in the Proximal Femur. Proceedings of the International Society for Magnetic Resonance in Medicine, 2002.

[7] E. Vardi-Gonen and G. T. Herman. Sequential vs. Simultaneous Stochastic Segmentation. IEEE International Symposium on Biomedical Imaging, pages 1327-1330, 2004. IEEE Catalog Number: 04EX821C, ISBN: 0-7803-8389-3.

[8] E. Vardi-Gonen and G. T. Herman. Sequential Binary Image Estimation. IEEE 31st Annual Northeast Biomedical Conference, pages 127-128, 2005. IEEE Catalog Number: 05CH3764C, ISBN: 0-7803-9106-3.

[9] E. Vardi-Gonen and G. T. Herman. Sequential Binary Image Estimation Algorithms. IEEE Signal Processing Society - 12th Digital Signal Processing Workshop; 4th Signal Processing Education Workshop, pages 494-499, 2006. IEEE Catalog Number: 06EX1488C, ISBN: 1-4244-0535-1.

Last Updated: 4/8/08