Dynamic Interfaces for Digital Libraries: The Open Video Project

1
Dynamic Interfaces for Digital Libraries The
Open Video Project

• Gary Marchionini
• University of North Carolina at Chapel Hill
• march_at_ils.unc.edu
• NJ/ASIST Distinguished Lecture
• April 4, 2002

2
Outline

• User Interfaces as crucial elements of DLs
• Overview of Open Video Project
• Agile Views interface framework
• Open Video examples of AVs
• User study example and near term plans
• Long term implications
• Summary

3
Observations

• Physical libraries architect space to aid
  information seeking
• Librarians interact with patrons in many ways
• Show a welcoming face
• Clarify needs and queries
• Assist with retrieval and use
• Digital libraries depend on the user interface to
  serve these purposes, thus user interfaces are
  crucial to success of high-volume digital
  libraries

4
Interface Principles for DLs

• Consider physical and conceptual interface
  issues.
• Consider representations and user control
  mechanisms.
• Provide a multiplicity of indexes
  (representations) Help people help themselves.
• Minimize user effort A click is a radical act.
• Use a variety of iterative user studies to
  understand user needs, common tasks, and
  interface effects.

5
Open Video Project

• Goals
• Create an open source DL for use by researchers,
  students, and the public.
• A testbed for interactive interfaces
• An environment for building a theory of human
  information interaction
• Ongoing work begun 1995 with colleagues at UMD
• Current funding NSF IIS-0099538, NCNI
• Collaborators I2-DSI, ibiblio, CMU, UMD, NIST,
  Internet Archive
• www.open-video.org

6
(No Transcript)
7
Current Status

• 0.5 TB of content
• 1600 video segments
• 1100 different titles
• 2000 unique visitors per month
• I2-DSI video channel
• OAI provider
• Ongoing user studies

8
Open Video Server
Distributed Files
Database (MySQL)
AVI
Search
Client (Browser)
MPEG etc.
Browse
Contribute
MPEG etc.
MPEG etc.
Digitization
Segmentation
Keyframe Extraction
Production System
Keyword (text)
Keyword (audio)
Surrogates
Metadata
9
Agile Views Interface

• Provide a variety of access representations
  (e.g., indexes) and control mechanisms
• Usual search and browse capabilities
• Leverage both visual and linguistic cues
• Create and test surrogates for overview and
  preview

10
Agile Views Framework
11
Evolution of Agile View Design Techniques

• Various dynamic query interfaces (HCIL UMD)
• Relation Browser (BLS, SILS seminars)
• Federal statistics, overviews of relationships
  (several different partitions). Useful for small
  number of attribute sets, each with small number
  of attribute values. Backend database of
  metadata, Java applet interface
• Enriched Links (Sony Labs)
• Complex web sites, previews, overviews, and
  reviews of pages. Backend computation and
  Javascript interface
• Integrated overviews and previews (BLC UMD)
• Multimedia digital library, backend computation,
  Java applet interface

12
Relation Browser
13
Enriched LinksPreview
14
Enriched Links Overview
15
Enriched Links Shared View
16
Overviews and Previews One Screen
17
The Open Video Project Case
18
Browse by Categories Attributes
Note Results Data
19
Search by Category Attribute
20
Search by Free Text Keyword
21
Search Results
Note Mouseover Popup for Details Click
yields Next segment Bib record
22
Segment Details
23
Video Transcript Text
24
Video Segment Preview
25
Research Agenda 2001-04

• What kinds of surrogates to provide for overviews
  and previews?
• Currently, we are designing and testing
  cost-benefit tradeoffs for
• Storyboards
• text keywords
• audio keywords
• Slideshows
• text keywords
• audio keywords
• Fast-forwards
• Integration of many specific good ideas lead to
  emergence or chaos?

26
AgileViews Overview Genre Documentary
Note One keyframe Per segment Shown
on Mouseover Direct select supported
27
AgileViews Overview Genre Education
28
AgileViews Overview Color/BW
29
Previews
30
Agile Views Preview Faces
31
Agile Views Preview Faces
32
Agile Views Preview Superimposition
33
Agile Views Preview Brightness
34
RD Scheme
35
(No Transcript)
36
User Study 1 Explore Surrogate Space (Fall 01)

• five video surrogates compared
• Storyboard (6x6 grid of keyframes) with text
  keywords
• Storyboard with audio keywords
• Slide show (keyframes displayed _at_ 250ms per
  frame) with text keywords
• Slide show with audio keywords
• Fast forward (at 4 times original speed)

37

• 10 subjects with video experience (about 2 hours
  each)
• Two phases
• Three surrogates for each of four videos
• Preferences changed over time, support for ff
  development
• Gist meant topicality, narrative structure, and
  visual style
• One surrogate (free choice) for each of 3 videos
• write statement of gist
• Select statements of gist
• Object recognition (textual)
• Object recognition (visual)
• Action recognition
• Visual gist

38
(No Transcript)
39
Next Steps (short term)

• Study 2 (spring 02)
• Compare different fast forward rates
• 30 subjects
• Interactive Shared Educational Environment (ISEE)
  for video (remote study)
• CHI 02
• The commons with demos and study of narrativity
• Summer study (02)
• Eye-tracking study of effects of textual and
  audio cues

40
Next Steps (long term) From DLs to the Sharium
41
Integration Hypothesis

• As information resources and technologies are
  integrated as digital libraries (sharia or
  collaboratories), institutional boundaries will
  blur. Examples
• Types of learning (formal, informal,
  professional)
• Types of libraries
• Levels of government (local, state, federal)

42
Long-Term Implications for Social Sciences
Research

• What does it mean when you can have everything
  you can possibly access anywhere, available
  everywhere? Removing the bounds of access
  implies ubiquity and augmented memory. What does
  it then mean to be informed? Intelligent?
  Consider a cascading set of issues such as trust,
  ownership/IP, communication, human relationships,
  and socio-technical symbiosis.

43
Summary Open Video as Testbed

• Give people many views to look ahead
• Make these views easy to manipulate (agile)
• Challenges
• Mapping video characteristics to surrogates
  (e.g., keyframes, keywords), mapping surrogates
  to control mechanisms (e.g., mouse actions)
• Automating production processes
• Use the repository, contribute to the repository!
  We would especially like to include cultural
  video in the collection.

44
Pointers and Thanks

• www.open-video.org
• www.ils.unc.edu/idl
• Thanks to
• National Science Foundation
• North Carolina Networking Initiative
• Contributors
• NJ ASIST