Transcript The Challenge of 3 Dimensional Data

Joint Conference on Digital Libraries, Portland, Oregon, July 15, 2002
The Challenge of
3 Dimensional Data Storage, Preservation,
Description, and Access
Dr. Jeremy Rowe
Head, Media Development, Information Technology
Partnership for Research in Stereo Modeling
Arizona State University
Partnership for Research In Stereo Modeling
(PRISM)
Established in Spring 1996 at Arizona State University to
promote interdisciplinary research in 3D Data Acquisition,
Visualization & Modeling, and Form Realization.
PRISM includes researchers with a shared interest in
visualizing 3D and higher dimensional data. Disciplines
include:
Anthropology
Archaeology
Biology
Computer Science
Educational Technology
Fine Arts
Industrial Engineering
Mechanical Engineering
Data Acquisition
Cyberware Laser Scanner
Also:
-Portable Laser Scanners
-Confocal Microscope
-CAT Scanner
-Other Method of 3D Point Cloud Capture
Laser Scanning to
Create Point Cloud
Single Scan
Multiple Scans Combine
to Create Point Cloud
Representing the Original Object
Surface Generation from Point Cloud
Assemble Point Cloud from Multiple Scans (Geomagic)
Point Cloud
Triangular Mesh Surface
3D Imaging Moving Beyond Viewable Images
Extending research capabilities beyond Image-Only
3D display such as QuickTime VR, Rhino, & Maya.
Funding - National Science Foundation Digital
Library/Knowledge and Distributed Intelligence
Goals to work with Domain Researchers to Develop
Processes to:
•Create A Quantifiable, Measurable Model
•Identify and Extract Data
•Create Catalog Information
•Populated Database
Surface Modeling
Fitting a Modeled, Measurable Surface
Spatial Measures of Similarity
Profile Curves
Graphically illustrate symmetry
And changes in curvature
Also:
Local Curvature
Inflection Points
Area
Volume
Proximity
Orientation
Etc.
Goal - to provide objective measures of objects and component features
Range of Projects
DNA in Fertilized Mouse Egg
 Through the Cell Biology Pilot
project, we applied computer
algorithms that enabled us to
obtain more information from
confocal microscope images.
Range of Projects
Native American Ceramic Vessels
 Digital collections to analyze
shape, morphology, manufacturing
techniques, improve measurement
accuracy, permit objective
comparison of objects, document
before repatriation.
Range of Projects
Bone Surfaces and Forensics
 Digital collections to analyze
shape, structure, morphology,
improve measurement accuracy,
permit objective comparison of
objects, and to identify unique
individuals.
Range of Projects
Feature Extraction
Morphology/Symmetry
Region Identification
Object Identification
Project Process Summary
Development of vocabulary
Development of XML schema and DTDs
Capture/digitization
Modeling and data extraction
Volume modeling research
Modeling software development
Data extraction
Data storage
Text and tabular data
Binary geometric/spatial data
Point cloud data
Polygonal mesh
Surface/volume models
Links to other databases
Query and interaction
Project Process Summary
Vocabulary
3D
3D Class
Class Inheritance
Inheritance
Each subclass definition inherits generic attributes from
its parent class and qualifies them or adds new ones
Class:Object
Attributes…
::Surface model
Shape
representation model
# features
Plus...
volume
surface area
::Vessel
Plus...
Axis
base plane
Diameter
circumference
::Jar
Plus...
Shoulder height
aperture diameter
Addition of 3D Data to
Relational Databases
Conventional Relational
Database Tables
Tables added to accommodate
3D Metadata
User Interface applications
(Java, Visual Basic, C++)
Database organization
3D Query
Processor
Existing Data
Access Protocols
(ODBC, JDBC,XML)
Text search
Attribute
Tables
RDBMS
(Oracle, SQL Server, etc)
Custom 3D
Data objects
(Surfaces, Volumes)
Spatial Search
3D Indexes
Query and Interaction
Creating a Digital Library
On Submit Query
1st CGI
1st JSP
Binary and Text
search criteria sent
from interface
Generate the contour type
and parameterize the text
query
Query sent
to
Databases
1st Record set
Database
Initial query interface
XSL/HTML
display
2nd Record set
2nd JSP
XML
document
Query
database
Pot ids of same
contour
2nd CGI
Curve
matching
X-Path/ display the selected pot
Query result with thumbnails
Detail display of individual pot
Examples of Current 3D
Data Applications
Environmental data
GIS
Air Quality
Water Quality
Built Environment
Building Design (CAD/Architecture modeling)
Building Documentation
Ground Penetrating Radar
Manufactured Objects
Design/Production/Quality Control
Catalog/Sale
Medical
CAT scans
Ultrasound
Data Components
Raw Data
Textual
2D Image/surface
3D Spatial
N-Dimensional
Time
Sensor data
Derived Data
Modeled Data
Technical Data
File Formats
Applications
Researcher Annotations (to permit reconstruction)
“Bookmark” Information Documenting Searches/Interaction
Metadata
• Traditional textual info (Dublin Core,
COPARES, etc.)
• Non-static information (audio, video, etc.)
• 2D data (from new and existing sources)
• 3D data (point cloud and modeled)
• Multi-dimensional data (i.e. vector, sensor,
time, etc.)
Class based XML Schema
*
Class based XML Schema
*
Class based XML Schema
*
Class based XML Schema
Other Issues
•Increasingly complex multi-dimensional data
•Development of standards
•File/storage formats
•Metadata/Cataloging/Description
•Researcher notations/version documentation
•“Bookmarking”/replicability of search results
•XML/SOAP search to replace/augment SQL search
•Migration
•Electronic records management of “Born Digital” data
3DK Project Team
Priti Aggarwal
Michael Bailey
David Capco
Gerald Farin
Mark Henderson
Adam Huang
Raghu Krovvidi
Mary Marzske
Pornchai Mongkolnam
S. Panchanathan
B. Ramakrishna
Jeremy Rowe
Arleyn Simon
David Van Alfen
Myungsoo Bae
D. Page Baluch
Daniel Collins
Patricia Green
Jiu Xiang Hu
Graham Jones
Dezhi Liu
Peter McCartney
Greg Nielson
Sandeep Pulla
Anshuman Razdan
Utsav Schurmans
Matt Tocheri
May Zsu
Contact Information
Partnership for Research in Stereo Modeling (PRISM)
at Arizona State University
http://3dk.asu.edu
Dr. Jeremy Rowe
[email protected]