Development of a number of successful techniques for large
population speaker recognition in addition to general research
and development in the field of pattern recognition with applications
to Speaker Recognition (Large-Scale Identification and
Verification), Segmentation, hybrid systems with face recognition
and text processing, large scale search techniques, etc.
Automatic Language Proficiency Rating
We have been successfully developing several
different techniques for the automatic rating of natural
human language proficiency in oral and written test
Research areas include all aspects of
Large-Vocabulary speech recognition, speech segmentation, and
Face Recognition and Detection
Research on face recognition and face detection. Most of our
research has included frontal portrait face detection and
recognition, to be used as a biometric.
Research areas include all control and recognition
activities through the use of Electroencephlography (EEG), but
concentrating mostly on signal and pattern recognition aspects.
Extensive work in this field including signature
compression and recognition (identification and verification).
Over 6 years of research in large-vocabulary, unconstrained
handwriting recognition. Extensive work on Feature
Extraction, Language Modeling, Compression, Segmentation and
Normalization leading to a fully operational system.
Information Theory and Language Modeling
Extensive research in Language Modeling and Information Theory
pertaining to usage with Text Processing, Handwriting Recognition,
Speech Recognition, and Hybrid Search systems using Textual Language
as well as Speaker Voice and Face information.
Applications of Optimization in very large Portfolio Optimization
problems, Learning Control, Adaptive Control, Neural Network
Learning. Very active in this field at the present including
Research done in conjunction with the Research department of
Adaptive Learning Control
Formulated for the first time the adaptive-learning control
strategies and provided solutions based on Optimization Methods,
Adaptive Ideas and new parameter estimation schemes. Also
formulated a continuous adaptive learning control system with a quite
strong and theoretically sound control strategy based on
previous work on adaptive parameter estimation and optimization.
Neural Network Learning
Break-throughs in highly efficient learning schemes using second
order techniques in optimization for Feed-Forward Neural Network
Systems. Several journal and conference papers in this field
including IEEE Best Paper Award for two conference
papers in the field.
Treatment of On-line Common Music Notation using a Pen.
Treatment of highly compressed dictionaries, statistical N-Grams,
Template Language models and Decision-Tree-based Language Models.
Created Loss-Less Image Compression schemes or storing highly
prized Art Collections for the Library of Congress.