C:/documents and settings/soma/my documents/resume/soma_cv/soma_resume_new.dvi

My research interest is in computer vision, image processing and pattern recognition. I have worked onillumination-invariant object representation and matching, shape indexing for matching and retrieval,2D face recognition, 3D face and object matching, fingerprint classification and related problems.
PhD in Electrical and Computer Engineering, University of Maryland, College Park.
August 2004 – PresentAdvisor: Prof. Rama Chellappa Masters in Electrical Engineering, Indian Institute of Technology (IIT) Kanpur, India.
August 2002 – May 2004 (GPA : 10.0/10.0)Advisor: Prof. Govind Sharma Bachelors in Electrical Engineering, Jadavpur University, Kolkata, India.
July 1997 – May 2001 (GPA : 3.8/4.0) • Albedo and Shape Estimation from a Single Image. We propose a non-stationary stochastic filtering framework for albedo and shape estimation using a Non-stationary Mean Non-stationaryVariance model for the true unknown albedo. The framework is extended to deal with imagesilluminated by multiple sources without having any information about the number or placement oflight sources.
• Articulation Invariant Shape Matching and Indexing. In this work, we develop an approach to index each shape based on a variety of simple and easily computable features that are invariantto articulations of part structures and rigid transformations. Shapes are retrieved using an efficientscheme that does not involve costly shape-wise alignment or correspondence establishment.
• Invariant Geometric Representation of 3D Point Cloud for Matching and Registration.
In this work, we develop isosurface-based representation to derive smooth and approximate char-acterization of input point clouds. Implicit function values on a set of suitably placed concentricspheres around the object are used as the features. Geometric invariance is achieved by sphericalharmonic decomposition.
• 2D and 3D Face Recognition. I worked on the Face Recognition Grand Challenge (FRGC) as part of the University of Maryland team. We developed and analyzed algorithms for matchingfaces across varying illumination conditions. For 3D face recognition, we use implicit surfacerepresentation that obviates the need for precise correspondence between surface points.
• A Non-generative Approach for Face Recognition Across Aging. Given the innumerable different ways in which a face can potentially age, it is very difficult to predict how a person willappear at a different age. In this work, we bypass the synthesis step and analyze the various agingeffects directly from a matching perspective. Our analysis is based on the observation that facialappearance changes in a coherent manner as people age.
• Role of Symmetry in Illumination-invariant Image Matching. We show that images are hardly ambiguous for the class of bilaterally symmetric Lambertian objects. The set of such objectscan be partitioned into equivalence classes such that it is always possible to distinguish between two objects belonging to different equivalence classes using just one image per object. Based onthe theoretical analysis, we propose a provably correct illumination-invariant matching algorithm.
• A Geometric Approach to Illumination Color Estimation and Specularity Removal in Images. In this work, we develop a purely geometric approach to estimate illumination chromatic-ity and separate diffuse and specular reflectance components from a single image. The approach isprovably correct for dichromatic surfaces and degrades gracefully when the dichromatic propertiesare not strictly satisfied.
Mentors: Dr. Nalini Ratha and Dr. Ruud Bolle (Exploratory Computer Vision Group).
• Exploring Curvature Information for Fingerprint Matching and Classification. In this work, we highlight the significance of high curvature points in fingerprints. We show that thecurvature information is reproducible and contains signature information important for the task ofmatching and classification. We provide experimental evidence that curvature-based classificationhelps in reducing the load of fingerprint indexing system without compromising on the accuracy.
• Fitting and Tracking in 2D and 3D Images Using Wavelet Based Deformation Model.
In this work, we define a probabilistic model that gives the prior distirbution for contour deforma-tion. The wavelet models are expressed in shape spaces for robusteness. The deformable modelis also used to generate a prior dynamical model for contour evaluation in time and tracking isperformed using Kalman filter.
Document Image Processing. The work involved dealing with challenges in automatic documentimage processing.
Responsibilties included taking lab classes for the following courses 1. S. Biswas, G. Aggarwal and R. Chellappa. Robust Estimation of Albedo for Illumination-invariant Matching and Shape Recovery. Accepted for publication in IEEE Transactions on Pattern Analysisand Machine Intelligence (PAMI), 2009.
2. N. Ramanathan, R. Chellappa and S. Biswas. Age Progression in Human Faces : A Survey.
Accepted for publication in Journal of Visual Languages and Computing (Special Issue on Advancesin Multimodal Biometric Systems), 2009.
3. S. Biswas, G. Aggarwal and R. Chellappa. An Efficient and Robust Algorithm for Shape Indexing and Retrieval. Under Review in IEEE Transactions on Pattern Analysis and Machine Intelligence(PAMI).
1. S. Biswas, G. Aggarwal and R. Chellappa. Robust Estimation of Albedo for Illumination-invariant Matching and Shape Recovery. In Proceedings of IEEE International Conference on Computer 2. S. Biswas, G. Aggarwal and R. Chellappa. Efficient Indexing for Articulation Invariant Shape Matching and Retrieval. In Proceedings of IEEE Computer Society Conference on Computer Visionand Pattern Recognition (CVPR), June, 2007.
3. G. Aggarwal, S. Biswas and R. Chellappa. Symmetric Shapes are Hardly Ever Ambiguous. In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition(CVPR), June, 2007.
4. S. Biswas, G. Aggarwal, N. Ramanathan and R. Chellappa. A Non-generative Approach for Face Recognition Across Aging. IEEE Second International Conference on Biometrics: Theory, Appli-cations and Systems, 2008.
5. S. Biswas, N. K. Ratha, G. Aggarwal and J. Connell. Exploring Ridge Curvature for Fingerprint In- dexing. IEEE Second International Conference on Biometrics: Theory, Applications and Systems,2008.
6. M. K. Johnson, D. G. Stork, S. Biswas, Y. Furuichi. Inferring illumination direction estimated from disparate sources in paintings: An investigation into Jan Vermeer’s Girl with a pearl earring.
SPIE Electronic Imaging: Computer image analysis in the study of art, vol. 6810, 2008.
7. S. Biswas, G. Aggarwal and R. Chellappa. Invariant Geometric Representation of 3D Point Clouds for Registration and Matching. In International Conference on Image Processing (ICIP), October,2006.
8. G. Aggarwal, S. Biswas and R. Chellappa. UMD Experiments with FRGC data. In IEEE Work- shop on Face Recognition Grand Challenge Experiments (CVPR), June, 2005.
9. S. Biswas and K. Nandy. Application of Wavelets in Detection and Classification of Microcalcifica- tion in Digital Mammograms - Some Recent Advances. International Conference on MathematicalBiology, 2004.
• Reviewer: 1) IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), 2) IEEE Transactions on Systems, Man, and Cybernetics - Part B, 3) IEEE Transactionson Information Forensics and Security (TIFS).
• Graduate student fellowship at University of Maryland (2004 - 2006).
• Merit Certificate for academic excellence fron Jadavpur University Alumni Association, 2001.
• B. N. Paul Memorial Gold Centred Silver Medal, Jadavpur University, 2001.
• Subodh Kumar Basu Memorial Medal, Jadavpur University, 2001.
Statistical and Neural Pattern Recognition • Dr. Rama Chellappa, Professor, Dept. of Electrical and Computer Engineering, University of Maryland, College Park, MD 20742. (rama@cfar.umd.edu) Niels Haering, Director, Government Science Development, Object Video, Reston, VA Nalini Ratha, Research Staff Member, Exploratory Computer Vision Group, IBM T.J.
Watson Research Center, Hawthorne, NY. (ratha@us.ibm.com)

Source: http://www.cfar.umd.edu/~soma/pdf/soma_resume.pdf