Dr. Kannappan Palaniappan

Our lab works on basic problems in bioimaging informatics and remote sensing using computational tools from image and video processing, computer vision, high-performance computing, scientific visualization, computer graphics, pattern recognition and real-time computing to help scientists and domain experts transform sensor data to information to knowledge. A current focus is on using level set methods for segmentation and object tracking of timevarying image and volumetric datasets.

A new frontier in the quantitative study of biological growth, motility, and morphogenesis is the use of molecular and cellular imaging (biomarkers) combined with advanced computational algorithms for high-throughput studies. Our NIH-supported contribution in this area including the four-color level set algorithm and directional geodesic active contours has enabled high-throughput automatic tissue and cell-tracking to quantify the variability of biological growth and differentiation at unprecedented spatial and temporal resolutions ranging from subcellular to molecular scales. New bioimaging methods being investigated include (epi-)fluorochrome, multispectral, multimodal imaging. Interactive flow visualization techniques will aid in the study and display of dense velocity volumes that can be correlated with biological processes. A variety of biological organisms, tissues and cells can be automatically analyzed to answer such questions as: How do cells find each other and sense their environment? How does an organism assemble itself without a wiring diagram? How do biochemical networks interact at the cellular level?

The Army Research Lab and Air Force Research Lab are supporting the analysis of wide-area motion imagery or large-format video that are on the scale of 16Kx16K and a few frames per second. Such wide coverage and large data volume pose significant computational and algorithmic challenges for the tracking of vehicles and people in dense urban settings. The ability to understand and classify the type of activity taking place in the scene leads to interesting computational tools for the analysis of high resolution spatio-temporal dynamics.

Non-rigid semi-fluid motion estimation for automatic deformable cloud tracking was a landmark NASA-funded study demonstrating the use of next generation satellite sensors and parallel computing (64K processors) for estimating dense wind vectors to improve numerical forecasting models and meteorological understanding of hurricane intensification. The study involved nearly a dozen researchers from two different government laboratories (NASA, NOAA) and several universities with data collection from satellite, aircraft and ground-based radar. The Massively Parallel Semi-Fluid Motion Analysis (MPSMA) algorithm, the first of its kind, demonstrated dense motion vector estimation using persistent differential geometric features with an accuracy of 1.5 m/s.

The Interactive Image SpreadSheet (IISS) and Distributed Image SpreadSheet (DISS) software environments were a pioneering research effort at NASA for visualizing and analyzing large image, multimedia, volumetric and information data collections in collaboration with A. F. Hasler and M. Manyin at Goddard Space Flight Center. The IISS was among the first tools to manipulate (hyper-)image and volume datasets by supporting rapid and smooth zoom, roam, animation, stereo viewing, formula execution and network-based data access, in synchrony involving any number of cells for arbitrary-sized multidimensional datasets. The IISS is a direct effort to meet the challenge of how the thousand-fold increase in data collected by NASA, occurring each decade, can be effectively used by a constant number of scientists. Terabyte and petabyte-sized digital multimedia libraries are being assembled to understand coupled systems (geophysical, biological, medical, astronomical, etc). The wide range of requirements for science, public outreach, and broad applicability of the approach to many imaging domains combined with the sustained improvements in computer technologies, need for ubiquitous access and opportunities for novel interfaces contributes to our continued research and development of the image spreadsheet visualization paradigm. It has been extended to applications in other domains including biomedical and multimedia. The extension of the IISS to the DISS (Distributed Image SpreadSheet) was the first NASA Earth Science application to demonstrate the feasibility of using high performance gigabit networks across country between desktops at NASA GSFC and the National Center for Atmospheric Research (NCAR) in Boulder, CO for remote interactive visualization applications. The DISS supports direct i/o, Web-based data access, uses data compression and network data caching to improve bandwidth utilization.

Panelist:
NIH Study Section Microscopy Imaging (MI), June 2008, Mehran Armand/Ross Shonat
NIH Study Section, ZRG1 BST-Q (01), a Special Emphasis Panel Continued Development and Maintenance of Biomedical Software, June 2008, Marc Rigas/George Chacko
NIH Study Section ZRG1 BST-R (40), P41 Resource Center: Electron Microscopy, Mark Ellisman, Univ of California-San Diego, March 2008, Ross Shonat
NIH Study Section ZRG1 BST-Q (01) S, Biomedical Software Maintenance and Extension, Feb 2008, George Chacko
NIH Study Section ZRG1 BST-E (50), Collaborations with National Centers for Biomedical Computing, 2007
NIH Study Section Microscopy Imaging (MI), 2007, 2006
NIH Study Section Biomedical Research Partnerships, 2006
NIH National Center for Research Resources P41, 2005
NIH Study Section ZRG1 BST-L (51), Continued Development and Maintenance of Biomedical Software, 2005
NIH Study Section ZRG1 BST-C (50), Data Sharing and Collaboration Tools, 2004
NIH Study Section ZRG1 SSS-H (90), Bioinformatics and Computational Biology Software, 2003
NIH Study Section ZRG1 SSS-H (91), Biological Software Maintenance, 2003
NSF Panelist, Information Technology Research (ITR), Directorate for Computer and Information Science and Engineering, 2003.
NSF Panelist, Strategic Technologies for the Internet, Directorate for Computer and Information Science and Engineering, 2002.
Grant Review, U.S. Civilian Research and Development Foundation (CRDF) Science Center programs, U.S. Department of State, 2002.

Conference Organization:
Program Reviewer, Int Conf Medical Image Computing and Computer Assisted Intervention (MICCAI), Oct 2008.
Program Reviewer, IEEE Int. Geoscience and Remote Sensing Symposium (IGARSS), Boston, MA, July 7-11, 2008.
Program Reviewer, Int Conf Medical Image Computing and Computer Assisted Intervention (MICCAI), Oct 2007.
Program Committee, 4th IEEE Consumer, Communications and Networking Conference (CCNC), Seamless Consumer Connectivity, Enabling Technologies Track, Las Vegas, Jan 2007.
Program Committee, 7th Asian Conference on Computer Vision (ACCV), Hyderbad, India, Jan 13-16, 2006.
Program Committee, Int. Workshop on Computer Architectures for Machine Perception, Palmero, Italy, July 4-6, 2005.
Program Committee, IEEE Workshop on Articulated and Nonrigid Motion (in conjunction with CVPR 2004), Washington, DC, June 27, 2004.
Program Committee, Int. Workshop on Computer Architectures for Machine Perception, Palmero, Italy, July 4-6, 2005.
Program Committee, Reviewer and Session organizer, IASTED Int. Conference on Biomedical Engineering, Salzburg, Austria, June 25-27, 2003.
Session chairman, IEEE Int. Geoscience and Remote Sensing Symposium (IGARSS), Information Systems, GIS and Visualization, Toronto, Canada, June 24-27, 2002
Session chairman, American Meterological Society Annual Conference, 18th IIPS, Orlando, FL, Jan. 13-17, 2002
Session chairman, American Meterological Society Annual Conference, 17th IIPS, Albuquerque, NM, Jan. 14-19, 2001
Program Committee, Reviewer and Session organizer, IASTED Int. Conference on Computer Graphics and Imaging, Las Vegas, Nevada, Nov. 20-23, 2000
Session chairman, IEEE Int. Geoscience and Remote Sensing Symposium (IGARSS), Visualization and Interactive Image Analysis, Honolulu, Hawaii, July 24-28, 2000
Session chairman, American Meterological Society Annual Conference, 16th IIPS, Longbeach, CA, Jan. 11-14, 2000

Reviewer:
IEEE International Symposium on Circuits and Systems
IEEE Transactions on Geoscience and Remote Sensing
IEEE Transactions on Signal Processing
IEEE Transactions on Pattern Analysis and Machine Intelligence
Computer Vision Graphics and Image Processing
International J. Imaging Systems & Technology
Int. Conf. Computer Vision and Pattern Recognition
Int. Conf.Pattern Recognition
Int. Conf. Computer Vision, Workshop onComputer Vision
NASA Graduate Fellowships in Global Change Research
NASA Center for Excellence in Space Data and Information Systems

The computational bioimaging algorithms being developed in my group can be extended and applied to other biological and biomedical applications such as cancer cell metastasis, tumor segmentation, vasculature segmentation, chemotaxis, deformable heart motion, multimodal 3D reconstruction, 4D confocal video microscopy, virtual endoscopy, etc.

Our current interests in ubiquitous sensor networks and large data visanalysis is in video tracking, geospatial event modeling, content-based multimedia retrieval using multi-core embedded realtime technologies, high density (wall) displays, immersive (VR/AR/multimodal) environments, wireless/wearable collaborative environments, etc.

Development of a New High Performance Computing Research Laboratory:
Recently we have developed a high-performance, reliable, low-cost, Linux-based cloud computing clusters that can be used for video and image mining applications. The Multimedia Communications and Visualization Laboratory (MCVL) established jointly with Dr. X. Zhuang for teaching and research consisting of $1.7 million of workstation hardware and software. MCVL has been funded by grants from the National Aeronautics and Space Administration (NAG-5-3900, NAG-5-6968, NAG-5-6283, NAG-13-99014), National Science Foundation (NSF 96-14 Award 9720668, NSF EIA Award 9911095), Raytheon (C-5-34366), NIH, NGA, and the College of Engineering. This hardware is used to support research in high performance visualization, virtual reality, image analysis and high performance networking.