Bijan Afsari
Email: bijan at cis dot jhu dot edu
I am a
postdoctoral research fellow in the Center for
Imaging Science (CIS) at the Johns Hopkins
University. I am with Prof. Rene
Vidal’s Vision Lab. Before that I was
with the Institute for Systems Research (ISR), University of Maryland. I obtained my Ph.D. in
Applied Mathematics under the direction of Professors P.
S. Krishnaprasad and Karsten Grove.
My Ph.D. and
current research are at the intersection of (Riemannian) geometry, statistics,
signal processing, numerical analysis, and control theory. At this time, I am
mostly focused on applications in computer vision and medical imaging. My
general interests include:
·
Applications of Riemannian geometry in engineering and computations.
·
Statistical signal processing.
·
Matrix joint diagonalization and tensor decomposition algorithms.
·
Nonlinear systems and data analysis.
·
Numerical analysis.
Some Publications:
- Bijan Afsari, ``A class of recursive-iterative means on
Riemannian manifolds,'' in preparation.
- Bijan Afsari, ``Means
and averaging on Riemannian manifolds,'' Ph.D. dissertation, University of
Maryland, College Park, November 2009 (available upon request).
- Bijan Afsari, ``Riemannian
$L^{p}$ center of mass: Existence, uniqueness, and convexity,'' Proc. Amer. Math. Soc. 139 (2011), 655-673.
- Bijan Afsari, ``A bias-variance dilemma in joint
diagonalization and blind source separation,'' in Proceedings of IEEE CAMSAP 2009, Dec. 2009, pp. 324-327.
- Bijan Afsari, ``What can make joint diagonalization
difficult?'' Proc. of ICASSP 2007, April 2007,
pp. III-1377-III-1380.
- Bijan Afsari, ``
Sensitivity analysis for the problem of matrix joint diagonalization'',
SIMAX Volume 30, Issue 3, 2008, pp. 1148-1171.
- Bijan Afsari, `` Simple LU and QR
based non-orthogonal matrix joint diagonalization'', Proceedings of ICA2006, Springer LNCS
series, 2006, pp 1-7.
- Bijan Afsari, P.S. Krishnaprasad,``
A
novel non-orthogonal joint diagonalization cost function for ICA'',
Institute for Systems Research Tech. Report, University of Maryland, 2005
- Bijan Afsari, P.S.Krishnaprasad,
``Some gradient based joint diagonalization methods
for ICA'', Proceedings of the Fifth International Conference on
Independent Component Analysis and Blind Source Separation, Springer LNCS series, 2004, pp. 437-444.
Note: If you are
interested in Matrix Joint Diagonalization please check here