Lalitha Agnihotri
E-mail: lalitha@cs.columbia.edu
Phone: (914) 945 6476
Education  
1999 - present Pursuing Ph.D. in Computer Science. GPA 4.0/4.0
1996 - 1998 Master of Science Computer Science and Engineering, August 1998.
The Pennsylvania State University, State College, PA. GPA 3.9 
1992 - 1996 Bachelor of Engineering Instrumentation & Control Engineering, June 1996.
Delhi Institute Of Technology, Delhi, India. 
Computer Skills * Three years experience as a C/C++ programmer
* System administrator for SGI ONYX, INDIGO and O2 workstations running IRIX
* Excellent knowledge of digital image processing and computer vision
* Working knowledge of UNIX shell programming

Languages: C/C++, Visual C++, Java, OpenGL, Perl
Platforms: UNIX (Solaris, IRIX), Windows (95/NT) 
Software Packages: Khoros/Cantata, MovieMaker, ShowCase, Word, Excel, PageMaker, Matlab

Experience  
08/98 – Present Philips Research, Briarcliff Manor, NY
Member Research Staff
Working on consumer information management project.
The goal of the project is to develop content representation and filtering technologies for image and video applications.
08/97 – 08/98 The Pennsylvania State University, State College, PA 
Graduate Teaching Assistant
Teaching undergraduate courses in Computer Vision in the Department of Computer Science and Engineering.
08/96 – 08/97 The Pennsylvania State University, State College, PA 
Graduate Research Assistant (Vision Group @ Penn State)
* Researcher in the field of Human Computer Interaction
* Administer SGI and Windows NT machines in the Computer Vision lab
06/95 – 08/95 Defence Science Center, New Delhi, India Summer Intern 
Designed and developed I/O interfaces for digital control of analog variables. Fabricated an ADC/DAC card.
Work@Philips Research Superhistograms: Implemented a technique "super-histogram" for video classification and program boundary detection. The image histograms in a non-linear color space are used to compare and merge sections of videos. 
Text Detection for Video Analysis: Developed and implemented a technique for detecting text in video frames. This is used for video analysis and for detection of commercials. VideoText has been accepted into MPEG-7 standards and is a part of the Final Draft International Standard.
Summarization: Developed and implemented technique summarizing and segmenting structured television programs using closed captions. The summary was later used for retrieval in a prototype we call Video Scout that watches television for people and brings the segments that a user might be interested in.
Projects Air Craft Scheduler: Designed an aircraft scheduler based on customer requests and aircraft availability. Temporal constraint propagation was used and domain specific knowledge was encoded into the system to be able to handle multiple aircrafts and multiple airports with multiple requests.
Face Recognition Using Neural Nets: An elementary face recognition system was implemented for recognizing faces. The system learnt the raw images using neural networks using back-propagation algorithms. When shown a new face, it classified them into one of twenty possible faces. The system could also detect the presence or absence of sunglasses on the face. 
Operating Systems: Developed and implemented JAVA program for controlling access to critical resources using synchronization techniques. Also implemented virtual memory's paging system using inverted page table.
Statistical Pattern Recognizer: Developed a statistical pattern recognition system for Gujarati script as a part of my graduate course work.
Histogram based image enhancement: Developed a Java applet for histogram based image enhancement techniques. This was used as a tutorial to introduce students to different image manipulation techniques used for enhancing the images.
Computer Vision Research: Developed and implemented a system for localizing and tracking a person's arm/hand in image sequences to be used further for gesture recognition. 
Speech Recognition: Designed a speech recognizer based on Hidden Markov Models (HMM's) using HTK toolkit.
Selected Publications N. Dimitrova, L. Agnihotri, G. Wei, Video Classification Using Object Tracking, Special Issue of the International Journal of Image and Graphics, Vol. 1, No.3, August 2001 (invited paper). 
T. McGee, L. Agnihotri, N. Dimitrova, "Transcript Triggers for Digital TV", Proceedings of 2001 IEEE International Conference on Multimedia and Expo, Tokyo, 2001.
L. Agnihotri, K. Devara, T. McGee, and N. Dimitrova, "Summarization of Video Programs Based on Closed Captioning", SPIE Conf. on Storage and Retrieval in Media Databases, San Jose, CA, January 2001, pp. 599-607.
C. Dorai, R Bolle, N. Dimitrova, L. Agnihotri, G. Wei, On the Evolution of VideoText Description Scheme and Its Validation Experiments for MPEG-7, Standards and Interoperability Workshop, ACM Multimedia 2000. 
N. Dimitrova, L. Agnihotri, C. Dorai, R Bolle, MPEG-7 VideoText Description Scheme for Superimposed Text. International Signal Processing and Image Communications Journal, September, 2000.
L. Agnihotri, N. Dimitrova, Video Clustering using superhistograms in large video archives, Visual 2000, Lyon, France, November 2000
L. Agnihotri, N. Dimitrova, Text Detection for Video Analysis, IEEE Workshop on Content-based Access of Image and Video Libraries, Colorado, Jun '99
Patents Over 10 patents filed.
References Available upon request.