Extending the Annotators Workbench

1
Extending the Annotators Workbench

• From EVIADA to CAMVA
• William G. Cowan
• Michael Durbin

2
History of EVIADA and CAMVA

• Ethnographic Video for Instruction and Analysis
  Digital Archive (started 2001)
• Extensive use of video from field collecting
• Desire to preserve this video
• Ability to annotate this video
• Networked access for research and instruction
• Central America and Mexico Video Archive (started
  2005)
• Goal Digitize, preserve, and make available
  hundreds of hours of at-risk or deteriorating
  film and video footage from three Central
  American archives
• Methodology Leverage tools and experience from
  EVIA Digital Archive project as much as possible

3
IU Partners for CAMVA

• Center for Latin American and Caribbean Studies
• Digital Library Program
• Center for the Study of History and Memory
• EVIADA

4
Central American Partners

• El Salvador
• MUPI Museo de la Imagen y la Palabra
• Nicaragua
• IHNCA Instituto de la Historia de Nicaragua y
  Centroamerica
• Mexico
• CIESAS Centro de Investigaciones y Estudios
  Superiore en Antropologia Socíal

5
CAMVA Support

• Grant from the Technological Innovation and
  Cooperation for Foreign Information Access
  (TICFIA) program of the US Department of
  Education
• October 1, 2005 September 30, 2009
• Project Director Jeff Gould, CLACS
• Project Manager Mike Grove, CLACS
• UITS Personnel Jon Dunn, Will Cowan
• EVIADA Support Mike Durbin, Shah Akram
• CAMVA Developer Gulshan Patil

6
CAMVA Overview

• Goal Digitize, preserve, and make available
  hundreds of hours of at-risk or deteriorating
  film and video footage from three Central
  American archives
• Methodology Leverage tools and experience from
  EVIA Digital Archive project as much as possible

7
CAMVA Process

• Video digitized by Central American institutions
  or local vendors
• Transmitted to IU via Internet or DVDs
• Transcoded to delivery formats at IU made
  available via IUs streaming servers and stored
  in MDSS / preservation repository

8
CAMVA Software Development

• Starting points
• EVIA Annotators Workbench
• EVIA Web Search Browse Interface
• EVIA Controlled Vocabulary Tool
• Operate within the larger technical environment
  at Indiana University
• Customize to be more appropriate to needs of CAMVA

9
Some differences between EVIA and CAMVA

• Need for multilingual interfaces for annotation
  and access
• Need for multilingual metadata, controlled
  vocabularies
• More varied genres of video
• Video being described by archivists or
  catalogers, not the original creator
• Less focus on preservation more on access

10
What we didnt want

• Too tightly coupled
• Need to meet the needs of CAMVA
• No flexibility
• Parallel Software Development
• i.e. CAMVA takes existing EVIADA code and does
  its own development
• Synchronization nightmare

11
What we decided to do

• Extend the Annotators Workbench
• Multilingual - start with Spanish but could be
  any language
• Flexible window layout - different windows and
  different layouts for different workflows
• Flexible data entry - different fields for
  different uses
• New XML Schema supports extended Annotators
  Workbench

12
Annotators Workbench B.C. (Before CAMVA)

• DEMO

13
Design and Implementation

• Internationalization
• Modularize the Annotators Workbench
• Determine scope of customizability
• Fields
• Components
• Features
• Create a means of expressing the configuration
• Layout XML Schema

14
Design Considerations Fields

• User interface must gracefully accept arbitrary
  inclusion of fields
• Largely manual process considering all possible
  combinations and looking at the user interface
• Entered metadata and annotation must be
  preserved, even when not exposed to the user in
  the current layout

15
Design Considerations Components

• A generalization of our component configuration
  must minimally support the current EVIA layout
  and the proposed CAMVA layout
• Components can be internal windows or anchored
• Components can have relative sizes or absolute
  sizes
• Components must fit into an arbitrarily sized
  container (the main application window)

16
Design Considerations Components

• Unable to find standard solutions to analogous
  formatting problems
• Predefined layouts for Java were unsuitable
• HTML doesnt deal with the same issues
• Doesnt seem to be any standard model for this
  kind of layout configuration

17
Design Considerations Components

• Final Design
• Anchored components are laid out sequentially in
  blocks that are anchored to one side of the
  remaining formatting area
• Floating components are placed in their
  percentage-based position
• Minimum sizes for floating component override
  layout specification

18
Demo of Annotators Workbench

19
Current and future work

• Video standards defined
• Developer hired (Gulshan Patil)
• Work begun on making Annotators Workbench
  multilingual
• Meeting with Member Institutions planned for
  Spring / Summer 2007
• Define user needs for Annotation Process
• Define user needs for Controlled Vocabulary
• Determine final digital preservation processes