LBNL Image Library -- Project plan

Image-Based Information Management and Access in the Scientific Environment

Create an LBNL-wide digital information management system for text, image, video, audio, and data objects, and multimedia combinations of these objects
Use this management system to organize and maintain the objects, and present the resulting information and knowledge to a wide audience
Provide operational capabilities adequate for a large, multidisciplinary scientific institution:
- Mechanisms that ensure physical integrity of corporate knowledge
- Both universal and controlled access to LBNL knowledge
- Long-term indexability (unique accession / object identifiers)
- "Natural" and convenient access across most platforms

Use the World Wide Web's proven ability to organize and present large amounts of diverse information that is managed by many different people
Specifically address the organization, storage, retrieval, and annotation of large collections of images by the scientific groups that generate those images (Initiate a sufficient level of dialogue to understand the wider issues and potential uses)
Build a prototype system around a small number of specific applications
Use mostly existing tools so that a prototype system can be operational in the near-term
Design a system that accommodates both text and image searching and browsing, as well as some "relational" capabilities

Direct involvement of the knowledge generators in the organization and maintenance of the information
An easily used mechanism for allowing multiple views of data
Wide availability of information (with controlled access) through the platform independent design of the WWW
Much better data integrity and access (through a "natural integration" with institutional tertiary storage (the LBNL Mass Storage System))

Provide simple mechanisms for owner specification of access control
Provide non-intrusive mechanisms for coupling auxiliary- and meta- data to archive data
Ensure that instantiated and non-instantiated referencing is easy and durable
Accommodate both text and image searching and browsing, as well as some "relational" extraction capabilities
Provide image and image meta-data browsing
Provide for "high speed" access (i.e. access limited by end-user workstation)
Accommodate derived data in a natural and consistent way, provide easy "pedigree" tracking for derived data
Provide mechanisms for maintaining user-application structured views of the data without incorporating this structure information into the basic storage scenario

Create a Web-based image management system that provides for:
- Extensible, annotated, modifiable, and easily accessed image archives
- "Easy" data-entry for images and annotations by data owners, digital librarians, analysts, etc. (e.g. scientific groups, Technical Photography, graphics artists)
- Fast access image storage facility
- Services that provide physical and logical integrity of information
Create a Mosaic-based environment that provides for:
- Image "browsing" and focused search capability based on the text annotations
- Retrieval and manipulation of image data and associated textual information
- (LBNL Image Library illustrates some prototype capabilities - see http://www-itg.lbl.gov/ImgLib/welcome.html)
Create tools and expertise that enable easily building and modifying the annotated image database
- Establish conventions for structuring image databases that will support scientific work groups
- Build tools that automatically populate these structures with original and derived image data, and the associated textual information
- Build tools that allow for modifying, augmenting, and updating this information based on owner defined access permissions (these tools should stay as close as possible to the WWW/Mosaic style of interaction for each of the supported platforms)
- Train digital librarians / digital photographers in the use of this technology
Build several example, production databases in order to refine and validate the approach