This research involves three research teams including State University of New York (SUNY) at Buffalo, National Center for Science Information Systems (NACSIS) at Japan, and the University of Nantes at France (IRESTE). The objective of the proposed research is to investigate novel approaches to supporting effective and efficient access to various geographic image databases over the Internet. These approaches will establish a foundation for the design of distributed geographic image retrieval systems. 

The technical challenge in the proposed research is the creation of a meta­level system on top of geographic image databases. The research issues to be addressed include the representation model for geographic data, the relationship between metadata and resource discovery on geographic data, and efficient query processing in a distributed environment. Thespecific research goals are: (1) investigate novel clustering approaches that can detect clusters of arbitrary shape of multidimensional data, which are  generated from geographic images, (2) construct a metadata model that formulates the metadata for  the integrated system to direct a query to relevant databases, (3) develop the theoretical foundation of  database selection approaches based on the metadata, and (4) design novel query processing approaches  that integrate heterogeneous queries extracted from the content of image data. 

Through the above research activities, the fundamental understanding of the relationships between  metadata and resource discovery on multidimensional data will be provided, and novel techniques will be  offered for designing a metaserver which integrates multiple geographic image databases. The expected  accomplishments from this research project include data resources and software tools. A huge volume of  geographic data will be collected and classified based on the semantics. This asset will be accessible by users  for research and educational purposes. A set of software tools will be developed and made available for  use of establishing a metaserver environment which manages large­scale multidimensional data resources.  The results obtained from this research is critical to retrieval on large volume of multidimensional  data distributed over Internet and will find broad applications as a template for the development of  systems which manage multidimensional data. Many applications, including financial, crystallography  and corporate databases, handle tremendous amount of multidimensional data. The approaches developed  in this research can be directly deployed in such environments. The experimental results to be generated  can be used to establish effective benchmarks for assessing the performance of distributed image data  retrieval systems. 


Results from Prior NSF Support 

Within the last three years, Zhang has been PI on three NSF grants. The first is a planning grant for women Scientists (IRI­9632394, $18,000, period 8/96--7/97). Second is a CAREER award (IIS­9733730, $200,000, period 5/98--4/02). Third is a CISE research instrumentation grant (EIA­9818289, $48,000, period 1/99--12/01). 

The first grant was provided for preliminary research into the impact of various quality­of­service(QoS) parameters on the multimedia presentation tools for education. The work is to identify those QoS parameters that are critical to the support of interactive learning and training activities and to conduct preliminary experiments with these parameters and candidate multimedia presentation methods. This preliminary research is an important precursor to the development of a full­blown research project on innovative multimedia presentation tools for digital libraries. A full­blown project based on the preliminary results resulted in the second CAREER award and the third instrumentation award. The design and experimental results from the above three grants include: (1) develop principles to support consistent multimedia presentations based on quality of service specifications; (2) develop adaptive end­to­end control techniques to support transmission of multimedia data over networks; (3) formulate a resource allocation model for a database server to allocate its resources for data retrieval; and (4) implementation and experimentation of a distributed multimedia database prototype for educational digital libraries. In the 3 years of these projects, Zhang wrote or co­authored a total of 36 items (including 11 journal articles, 1 edited journal special issue, and 24 articles in conference and workshop proceedings). She served as primary advisor of 2 doctoral degree recipients and 17 master's degree recipients, all of whom had significant links to the project. 

Zhang is also a senior participant on a five year, $2.2 million IGERT award, ''Integrative Graduate Education and Research Training in Geographic Information Science'' (DGE 9870668) with Mark (PI). Within the last five years, Mark has been PI or co­PI on four NSF grants or their subcontracts. The first was a grant (SBE 9412901) to support Ann K. Deakins doctoral research. Second was the award to support the National Center for Geographic Information and Analysis (NCGIA; SBR 8810917); Mark was PI of the Buffalo subcontract of that award. Third, Mark is PI of the Buffalo subcontract of the award ``Advancing Geographic Information Science'' (SBR 9600465; known as ``Project Varenius'') for the period 1997­2000. In the Varenius Project, Mark chairs the panel on Cognitive Models of Geographic Space, which oversees three specialist meetings in 1998­99 (http://www.ncgia.org/varenius). Fourth, Mark is PI and Zhang is Senior Personnel on a five year, $2.2 million IGERT award, ``Integrative Graduate Education and Research Training in Geographic Information Science'' (DGE 9870668). 

The project most relevant to the current proposal is the award ``National Center for Geographic In­ formation and Analysis'', SBR 8810917, awarded for the period 1988­1996, $9,800,000 (all sites). Selected results of this project are given below. Full documentation is available in the Annual Reports of NCGIA, available as part of the NCGIA publication series at http://www.ncgia.ucsb.edu/pubs/publications.html. 

National Science Foundation award SBR 8810917, which established the NCGIA, was granted in 1988 to a consortium consisting of the State University of New York at Buffalo, the University of California at Santa Barbara, and the University of Maine. Mark co­led the NCGIA's Research Initiative 2, ``Languages of Spatial Relations,'' (Mark and Frank 1992), Research Initiative 13, ``User Interfaces for Geographic Information Systems'' (Mark and Frank 1996), and Research Initiative 21, ``Formal Models of Common­ Sense Geographic Worlds'' (Egenhofer and Mark, 1995b). In the 8 years of the project, Mark wrote or co­authored a total of 73 items (including 9 journal articles, 2 edited books, 14 book chapters, 34 articles in conference proceedings, and 14 technical reports); these contributed to 8 of the 19 NCGIA research initiatives. He served as primary advisor or chaired the committees of 6 doctoral degree recipients and 10 master's degree recipients, all of whom had significant links to the project. 

Under NCGIA Research Initiative 10 (``Spatio­Temporal Reasoning and GIS'') Mark worked with Max Egenhofer (Spatial Information Science and Engineering, University of Maine) on a series of human subjects experiments to confirm and refine a mathematical model of spatial relations. Testing was done with subjects from five natural languages, and the cross­linguistic testing of the formal computational model led to new insights both into the formal model itself and into the ways people use locative expressions in the domain tested. More than 600 subjects were tested using three different protocols (Mark et al. 1995). This research produced 12 papers written or co­authored by Mark, of which the most important are Mark and Egenhofer (1994a, 1994b, 1995) and Egenhofer and Mark (1995a).


Research Initiative 21 (``Formal Models of Common­Sense Geographic Worlds) was co­led by Mark. The goals and objectives of this Initiative were to identify basic elements of common­sense conceptualizations of geographic entities and processes, to develop an integrating framework within which the relations between these elements can be expressed, to investigate different types of intuitive geographic inferences, and to compare these inferences with current GIS technology. The Specialist Meeting for this Initiative was held in fall 1996 in San Marcos, Texas, and a report summarizing its results has been published (Mark et al. 1997). In one Initiative 21 project, Mark has been working toward establishing an ``ontology of geographic entities'' and testing the cross­cultural universality of such an ontology. 

The IGERT grant (DGE 98­70668) began in September 1998, and will admit its first Doctoral students in fall 1999. This program of ''Integrative Graduate Education and Research Training in Geographic Information Science'' will expose Doctoral students in any of seven participating academic departments (including Geography and Computer Science) to research in any of 6 research areas. One of these research areas is labeled ``Cognitive Models of Geographic Space.'' This topic is led by Mark, and has geographic ontology as the major theme. It is likely that some of the IGERT­funded students will gain research experience by participating in the present proposed project (at no charge to this project) if it is funded. 

Acharya's recent NSF grant include ``Virtual Real Time Visualization Laboratory'', $111,714 (With SUNY Matching), PI: Kesh Das, Co­PIs: Raj Acharya, C. Bloebaum, R. Nagi. This grant involves developing a real­time visualization laboratory. Raj Acharya is involved with three aspects of research using the laboratory: (a) Spatio­temporal Visualization of multimedia databases, (b) Visualization of multidimensional microstructures, and (c) Virtual reality visualization of biomedical structures. Acharya has published 11 journal and conference papers relating to the above project.