Assistant Computer Science
Resume:
Ajit Rajwade
Computer and Information Science and Engineering Department
CSE Building, Room E301,
P.O. Box 116120,
Gainesville, FL 32611-6120, USA.
***@****.***.***
********@*****.**.**
http://www.cise.ufl.edu/~avr
Education University of Florida, Gainesville, Florida, USA.
Ph.D. in Computer Science, Jan. 2005 to Oct. 2010, GPA: 3.87
Advisors: Anand Rangarajan, Arunava Banerjee.
Thesis title: Probabilistic Approaches to Image Registration and Denoising.
McGill University, Montreal, Quebec, Canada.
M.Sc. in Computer Science, Aug. 2002 to Dec. 2004, GPA: 3.72
Master s thesis title: Facial Pose Estimation and Face Recognition from 3D Data.
Government College of Engineering, Pune (University of Pune), India.
B.E., Aug. 2001, Computer Engineering (First class with distinction in every semester ).
Areas of Image/video restoration: currently working on denoising, interested in removal of com-
Research pression artifacts, deblurring, dejittering, removal of fog or haze, contrast enhancement,
Interest relighting, inpainting, dither.
Image and video compression using machine learning and sparse representations, inpainting-
based methods.
Probability density estimation.
Biometrics, especially face recognition.
Problems in computational geometry (area and perimeter based skulls and hulls).
CAPTCHAs and Automated Turing Tests.
Best Scienti c Paper Award at the International Conference on Pattern Recognition (ICPR),
Awards/
2008, Tampa, Florida.
Honors
Selection with full funding at Summer School on Image Processing at the Park
City Math Institute/Institute for Advanced Study (IAS), Park City, Utah in June/July
2010, [declined].
Travel Grants from the Department of CISE at the University of Florida to present pa-
pers at the Workshop on Energy Minimization Methods in Computer Vision and Pattern
Recognition (EMMCVPR), Bonn, Germany (Summer 2009), and from the Department of
CISE and the Student Government at the University of Florida to present my paper at
the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), New York
(Summer 2006).
Tuition Fee Waiver for Summer 2003, McGill University (awarded to two international
students out of 100 in the year 2003-2004, in recognition of academic excellence).
Rank 8 in Computer Engineering at the University of Pune (out of 800 students) in
the year 1999-2000.
Ajit Rajwade
Image Filtering
Research/
Project description: This project includes
Publications
1. two novel techniques to learn spatially adaptive transform bases for image ltering,
building upon the singular value decomposition of the image patch (the methods pro-
duce close to state of the art results, are simple to implement and have parameters
tuned in a principled way).
2. a new hypothesis-testing based criterion for automatically selecting the optimal pa-
rameters of a lter to denoise images corrupted by i.i.d. noise,
3. a technique for ltering driven by low-frequency interpolation (and resulting isocon-
tours) between the pixels of a noisy image,
Journal papers in preparation.
Ajit Rajwade, Anand Rangarajan and Arunava Banerjee, Automated Filter Parameter
Selection using Measures of Noiseness, Canadian Robot Vision Conference, June 2010,
Ottawa (Canada).
Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, Image Filtering Driven by
Level Curves, International Conference on Energy Minimization Methods in Computer
Vision and Pattern Recognition (EMMCVPR), Aug 2009, Bonn (Germany). [Oral
Presentation]
Image Compression
Project description: A technique for compression of images (or image databases) by learning a
group of orthonormal matrix pairs to compactly represent small-sized patches. The learning
technique exploits patch-level redundancy within and across images.
Karthik Gurumoorthy, Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, A
Method for Compact Image Representation using Sparse Matrix and Tensor Projections
onto Exemplar Orthonormal Bases, IEEE Transactions on Image Processing (TIP),
Volume 10, Issue 2, Feb. 2010, pp. 322-344.
Karthik Gurumoorthy, Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, Be-
yond SVD: Sparse Projections Onto Exemplar Orthonormal Bases for Compact Image
Representation, International Conference on Pattern Recognition (ICPR), Dec 2008,
Tampa (USA). Best Scienti c Paper Award.
Probability Density Estimation for Image/Signal Processing
Project description: This is a new technique for estimating the probability density of the
values of a signal (image), driven by the choice of an interpolant to approximate a digital
signal as a continuous function. This results in beautiful relationships between geometric
entities such as the gradients or level curves of an image and the probability density function
of the intensity values. Extensions to gradient-biased density estimation and joint density
estimation for multiple images are also presented. The method is not based on external
kernels as in classical kernel density estimation.
Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, Probability Density Estima-
tion using Isocontours and Isosurfaces: Applications to Information Theoretic Image
Registration, IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI),
Volume 31, Issue 3, March 2009, pp. 475-491. (This work has been cited by various re-
search groups in computer vision, visualization and medical imaging, and in two recent
books: Information Theory in Computer Vision and Pattern Recognition by F. Escolano,
S. Pablo and B. Bonev, (September 2009) and Image Fusion Theories, Techniques and
Applications by H. Mitchell (March 2010), both published by Springer Verlag )
Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, A New Method of Probability
Density Estimation with Application to Mutual Information Based Image Registration,
Ajit Rajwade
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2006,
New York (USA) [Oral presentation, acceptance rate 6 %].
Ajit Rajwade, Arunava Banerjee and Anand Rangarajan, Continuous Image Repre-
sentations solve the Histogram Binning Problem in Mutual Information Based Image
Registration, International Symposium on Biomedical Imaging (ISBI), April 2006, Ar-
lington, VA (USA).
Partially Observable Markov Decision Processes
Masoumeh Tabaeh Izadi, Ajit Rajwade and Doina Precup, Using core beliefs for point-
based value iteration, International Joint Conference on Arti cial Intelligence (IJCAI),
August 2005, Edinburgh (UK).
Face Recognition
Ajit Rajwade and Martin Levine, Facial Pose from 3D Data, Image and Vision Com-
puting, Volume 24, Issue 1, August 2006, pp. 849-856.
Research Research assistant at the University of Florida Work on probability density estima-
Work and tion of image intensity values (applied to information theoretic image registration), image
Employ- compression and image denoising (January 2005 to present).
ment Intern at GE Research, Bangalore, India in the Imaging Technologies Group. Worked
History on level-set based registration of MR volumes (Summer 2007).
Research assistant at McGill University Department of Computer Science, working
on Partially Observable Markov Decision Processes (POMDPs) (Fall 2005).
Research assistant at McGill University Department of Agriculture, applied machine
learning techniques to determining nitrogen content from hyperspectral features of soil (Fall
2005).
Thesis on 3D face recognition at McGill University (Summer 2003 to Fall 2004).
Associate Software Engineer at Veritas Software India Ltd., now part of Symantec Corp.
(August 2001 to August 2002).
Teaching Served as course instructor for:
Experience a course on Applications of Discrete Structures at the University of Florida (Dept.
of CISE), for a class of 110 undergraduate students.
a course on Calculus instructor for a class of 20 students, as part of the University of
Florida STEPUP program for smooth transitioning of minority high school students
into university life.
Served as teaching assistant for:
Analysis of Algorithms (Graduate Course), UF, Spring 2010.
Data Mining (Graduate Course), UF, Spring 2009.
Applications of Discrete Structures, held discussion sessions for undergraduate
students for totally 5 semesters.
Courses at McGill University: Programming Languages, Theory of Computer Science,
Introduction to Java Programming, and Data Structures and Algorithms, Fall 2002 to
Summer 2004.
Served as Guest Speaker for:
Lectures on information theory and constrained optimization to a class of graduate
students for the course Mathematical Methods for Intelligent Systems, Fall
2006.
Ajit Rajwade
Lectures for the course Advanced Machine Learning on probability density esti-
mation and its application in medical imaging, Fall 2008.
Lectures for the courses on Applications of Discrete Structures and Numerical
Analysis on the topics of recursion and recursive function de nition, and root- nding
techniques respectively, Fall 2008.
Talks At CVPR 2006, New York A New Method of Probability Density Estimation with
Application to Mutual Information Based Image Registration .
At UF Landmines Seminar, October 2008 Probability Density Estimation using Iso-
contours and Isosurfaces: Applications to Information Theoretic Image Registration .
At EMMCVPR 2009, Bonn, Germany Image Filtering Driven by Level Curves .
Reviewer for following journals:
Professional
Service IEEE Transactions on Pattern Analysis and Machine Intelligence for the years 2005,
2006, 2007 and 2010.
IEEE Transactions on Medical Imaging, 2008 and 2009.
Machine Vision and Applications, 2006, 2009.
Computer Vision and Image Understanding, 2007, 2008 and 2010.
Reviewer for following conferences:
CVPR, 2009.
MICCAI, 2008.
EMMCVPR 2005.
Seminar Coordinator for CVGMI (Computer Vision, Graphics and Medical Imaging)
research group at the CISE department, University of Florida, Fall 2005 to Fall 2006, and
Fall 2008.
Student Volunteer for EMMCVPR 2005.
Student Member of IEEE since 2005.
Relevant At University of Florida Computer Vision, Machine Learning, Medical Image Analysis,
Course- Advanced Topics in Computer Vision, Advanced Topics in Machine Learning, Computer
work Graphics, Linear Algebra, Numerical Optimization, Combinatorics, Computational Geom-
etry, Analysis of Algorithms, Algebraic Geometry, Formal Languages.
At McGill University Computational Perception, Shape Analysis, Image Processing, Sta-
tistical Computer Vision, Robotics, Reinforcement Learning, Arti cial Intelligence, Com-
putational Geometry.
At University of Pune, India Digital Signal Processing.
Skills Languages C (expert), C++ (advanced), MATLAB (expert), Java (basic).
Operating Systems Linux, Windows.
Packages GNU Scienti c Library (GSL), Visualization Toolkit (VTK) and Insight Toolkit
(ITK).
Hobbies President of the UF chapter of SPICMACAY (Society for Promotion of Indian Clas-
sical Music and Culture Amongst Youth), Summer 2009 to Summer 2010.
Writing English poetry
Learning languages (German, French) besides English, Hindi and Marathi.
Ajit Rajwade
Playing keyboard (elementary)
Yoga
Ardent listener of all forms of Indian music
Member of magazine committee at the Government College of Engineering in the year
1998-1999.
References Available upon request
de2.dvi
Contact this candidate