Research Achievements and Continued Projects

My basic expertise is in adaptive learning, medical imaging, intelligent image analysis, data security protocols and neural networks. I have applied my research work in many interdisciplinary applications and have achieved local, national, and international recognition. I received two US patents for my inventions: The Nevoscope for Imaging Skin Lesions and the Neural Network Based Image Enhancement System. I received the Martin Epstein Award in 1984, the prestigious NIH young investigator F.I.R.S.T. Award in 1988, the ECE departmental "Distinguished Progress in Research Excellence Award" in 1989, the "Sigma-Xi Young Investigator Award" in 1992, the UC "Faculty Achievement Award" in 1994, the "IEEE Engineering in Medicine and Biology Early Career Achievement Award" in 1995, and the Dorman Distinguished Lectureship Award in 1999 for contributions in medical imaging and intelligent image analysis. I also established an interdisciplinary! research group in Simultaneous MR-PET Medical Imaging at the UTSW Medical Center and UTA. At NJIT, I founded the NJ Center of Wireless Networking and Internet Security. Some of my achievements are listed as follows.

  1. Invented a new imaging instrument, called the "Nevoscope" for imaging skin-lesions for detecting fatal skin-cancer melanoma. Three-dimensional image reconstruction and analysis algorithm were developed to extract important diagnostic and prognostic features from the image yielded by the Nevoscope. This research work was selected as one of the three Scientific Exhibits under the category of Original Investigation presented at the 48th Annual International Meeting of the American Academy of Dermatology held in San Francisco from Dec. 2-7, 1989. This patented technology is now being manufactured and marketed by Translite Inc. and being evaluated at five medical centers in the nation. Recently, I have developed a Multispectral Nevoscope for which clinical evaluations are being pursued. A second US Patent was filed for the Multispectral Nevoscope in 2000.

  2. Developed new feature enhancement and feature extraction algorithms for analyzing microcalcifications of the breast mammograms using wavelets and texture analyses. These features are then used in an artificial neural network based classification system for detecting breast cancer. I was invited to present my research work in the International Conference of American Association of Physicists in Medicine held in August 1992 in Calgary, Canada. I authored a review paper on the mammographic image analysis in a clinical journal "Mammo-Matters".

  3. Developed a Multi-Grid Expectation Maximization (MGEM) image reconstruction algorithm for the PET data. The algorithm provides the Maximum Likelihood estimation using the multi-resolution pyramid, and has produced much better and computationally more efficient results than the conventional single-grid Expectation Maximization algorithm.

  4. Developed a new Iterative Principal Axes Registration algorithm to correlate three-dimensional multi-modality medical images. A knowledge-based analysis system has been developed for interpreting the registered Magnetic Resonance and Positron Emission Tomography images of brain. This work has attracted national and international attention and funded by Hoechst Pharmaceuticals. Patient data from UC and the Kettering Medical Centers have been analyzed in my laboratory. Physicians are now using this analysis for diagnosis. The Ph.D. dissertation of Louis Arata was completed on this work. I chaired the International Symposium on Knowledge-Based Multi-Modality 3-D Medical Image Analysis in the IEEE Annual International Conferences of Engineering in Medicine and Biology in Philadelphia in 1990, in Orlando in 1991, and in Paris in 1992.

  5. Developed new neural networks based models for non-linear function approximation for applications in fault diagnosis and isolation system of the Space Shuttle Main Engine. This work is funded by NASA Space Engineering Center and is being done in collaboration with the NASA Lewis Research Center. The collaboration with Aerojet and United Technology Research Center has been established to develop SSME sensor validation system using neural network based models of SSME parameters.

  6. Developed new recombination operators for use in Genetic Algorithms. Based on the Global Random Search methods, new Genetic Algorithms have been developed to obtain high-performance samples in a large parameter search space. These algorithms with new recombination operators are shown to converge to the global optimum solution in a noisy environment. The Ph.D. dissertation of Charles Peck was completed on this work.

  7. Developed a new Chord-Tangent Transformation for recognition of target objects in a noisy environment with occlusion. This transform is 100 to 200 times more efficient than the Generalized Hough Transform which is considered to be the standard technique in target recognition. The Chord-Tangent Transformation is particularly useful in robotic vision and defense related applications. The Ph.D. dissertation of Thomas Dufresne was completed on this work.

  8. Developed a new optimal Morphological Signature Transform (MST) for shape description, representation, and matching. The MST is based on multi-resolution non-linear morphological processing of a given shape using a structuring element. The design of the structuring element is optimized using a Genetic Algorithm. The MST has been shown to be robust against the noise and non-linear distortions. The Ph.D. dissertation of Sven Loncaric was completed on this work.

  9. Developed the "Knowledge-Based Computer Vision" laboratory and the Center for Intelligent Vision and Information Systems at the University of Cincinnati for conducting research in image analysis and interpretation for both medical and industrial applications.

  10. Developed strong collaborating links between the Department of Electrical & Computer Engineering and the Departments of Radiology, Neurology, Dermatology, and Molecular Genetics of the UC Medical School. I was deeply involved in initiating an interdisciplinary biomedical research program between the College of Engineering and the College of Medicine at the University of Cincinnati. I successfully established the collaborative research partnerships with the Kettering Medical Center, Thomas Jefferson Hospital, Brookhaven National Laboratory and the Texas Medical Center at Houston. Collaborative research proposals were submitted to the NIH. In addition, collaborations with several industry partners including MTL Systems, NASA LeRC, Westinghouse Science and Technology Center and Procter and Gamble were successfully established with my research laboratory.

  11. I am a part of the consortium with NASA Glen, NASA Ames, Pratt & Whitney and United Technologies on Data Mining and Optimization Tools for development Aviation Safety System. This work is continued for the development of an Adaptive Learning System for Event Detection and Characterization (ALEC) using neuro-fuzzy systems.

  12. Recently, I established the New Jersey Center for Wireless Networking and Internet Security at NJIT in collaboration with Princeton University and leading industries. The Center focuses on collaborative research to develop future technologies in network management and security protocols for 4th generation wireless networking systems. I am developing a multi-level dynamic security protocol using wavelet transforms and watermarking technologies.