Introduction
Through my undergraduate studies at Caltech and research experiences at Los Alamos National Laboratory and Lincoln Laboratory, I have developed broad interdisciplinary interests in the fields of signal and image processing, artificial intelligence, and distributed sensor networks. These diverse studies and research experiences have fueled a keen interest in problems related to computer vision – a topic I was first introduced to as an undergraduate student at Caltech. To further my development as a computer vision researcher, I am currently pursuing a doctorate in Electrical Sciences and Computer Engineering at Brown University.
As a graduate student at Brown, I am investigating solutions to problems in computer vision utilizing networks of distributed smart cameras. I am principally interested in (1) developing novel systems for the acquisition of three-dimensional representations of objects and scenes from multiple-viewpoints and (2) investigating techniques that enable the rapid processing of visual information – particularly embedded systems and distributed network architectures. At a fundamental level, I am most interested in understanding how humans extract meaning from visual scenes and in developing artificial systems that realize similar or complementary capabilities.
Previous Research Experience: Computer Vision
My first exposure to signal and image processing occurred as an undergraduate student at the California Institute of Technology, where I completed a project-oriented course on 3D photography offered my junior year by Prof. Gabriel Taubin, then on sabbatical from the IBM Thomas J. Watson Research Center.
As a final project for Prof. Taubin's course, I worked with two other undergraduates to develop a novel system for capturing photo-realistic 3D models from a series of photographs of a human subject. Starting with front and profile views, we employed a user-assisted procedure to recover 3D coordinates of a sparse set of feature locations on a subject's face. A scattered data interpolation algorithm was then applied to deform a generic model to fit the particular facial geometry. Afterwards, a model texture map was created by combining both original photographs, producing a photo-realistic three-dimensional model.
Motivated by my brief exposure to computer vision as an undergraduate, I accepted an assistant staff position at MIT's Lincoln Laboratory in the summer of 2002. At Lincoln, I studied numerous methods for object recognition, image restoration, and parameter estimation, as well as developed detailed simulations of sensor-specific image formation processes for ballistic missile seeker and interceptor applications.
Directed by Dr. John-Scott Smokelin while at Lincoln, I investigated numerous image processing algorithms including: Hough transform methods, corner detection schemes, morphological operations, and moment-based sub-pixel edge detection. These studies were documented in Advanced Algorithms for Endgame Aimpoint Selection, published in the Proceedings of the First Missile Defense Conference held in Washington , DC in March of 2003.
In addition to these studies, Dr. Smokelin and I investigated novel methods for image restoration including wavelet-based regularized deconvolution. One of our primary accomplishments was successfully demonstrating a unique algorithm to synthesize a single super-resolution image from a sequence of blurred, noisy, and under-sampled images. Our image restoration studies were documented in Seeker Super-Resolution and CSO Detection, published in the Proceedings of the 2003 Military Sensing Symposium Specialty Group Meeting on Missile Defense Sensors, Environments, and Algorithms held in Monterey, CA in November of 2003.
Through my employment at Lincoln Laboratory, I developed a more complete understanding of signal and image processing, establishing what I believe to be the necessary technical foundation for my current graduate studies in computer vision.
Previous Research Experience: Sensor Networks
As a junior at Caltech, I completed an additional project-oriented course focusing on topics in the then-emerging field of Swarm Intelligence. As a final class project, I worked with graduate student William Agassounon to investigate a labor division mechanism for mobile robotic platforms which mimicked the behavior of social insects. My contributions to the problem were acknowledged in A Scalable, Distributed Algorithm for Allocating Workers in Embedded Systems, published in the 2001 Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics.
In order to more fully explore my interest in distributed sensor networks, I completed an internship at Los Alamos National Laboratory in the summer of 2001. My principle research at Los Alamos involved Distributed Sensor Networks with Collective Computation (DSN-CC), also referred to as ad-hoc wireless sensor networks. Working in close collaboration with Dr. Anders Jorgensen, I developed a general simulation environment to study the collective behavior exhibited by sensor networks. In addition, I formulated a new algorithm for fully distributed audible source localization derived from the Time Difference of Arrival (TDOA) method.
Conclusion
Through my research experiences at Caltech, Los Alamos, Lincoln Laboratory, and recently at Brown, I have contributed to numerous research projects in diverse fields extending from computer vision to collaborative robotics – demonstrating both my academic capabilities and my determination to pursue advanced studies in emerging technologies. Motivated by my passion to conduct original research, I am currently enjoying continuing my interdisciplinary studies as a graduate student in Electrical Sciences and Computer Engineering at Brown University.
Last Updated: February 17, 2009