Nicky Ransom

1

• Nicky Ransom
• University for the Creative Arts

2

• Tags
• William Hall Raine
• crowd
• men
• hats
• street
• night
• lighting
• faces
• sea of people
• people
• watching
• event
• election
• results
• populated

Election night crowd, Wellington,
1931 Photographer William Hall RaineElection
night crowd, Wellington, 1931Reference number
1/2-066547-FOriginal negative Photographic
Archive, Alexander Turnbull Library
3

• Growth in number of images online
• Accurate and comprehensive indexing is critical
  to make online content accessible
• But visual materials are difficult to index

4

• Concept-based indexing assigning index terms to
  describe the subject of an image

5

• Concept-based indexing assigning index terms to
  describe the subject of an image
• Search engine indexing index terms
  automatically created from data related to an
  image

6

• Concept-based indexing assigning index terms to
  describe the subject of an image
• Search engine indexing index terms
  automatically created from data related to an
  image
• Content-based indexing using automatic
  processing to index image attributes such as
  colour, texture and shape

7
CIRES Content Based Image REtrieval System
8
(No Transcript)
9

• To find out value of tags for image retrieval by
  investigating whether the terms used to describe
  images in tags are similar to the terms used to
  search for images.
• Which image facets are described in user tags?
• How do these compare to those found in image
  queries?
• What are the implications for future use of
  tagging for online indexing?

10
Armitage, L., Enser, P. (1997). Analysis of
user need in image archives Journal of
Information Science, 23(4), 287-299.
Specific Generic Abstract
Who? Individually named person, group or thing (S1) eg Napoleon Kind of person, group or thing (G1) eg Skyscraper Mythical or fictitious being (A1) eg King Arthur
What? Individually named event or action (S2) eg London Olympics Kind of event, action or condition (G2) eg Football game Emotion or abstraction (A2) eg Anger
Where? Individually named geographical location (S3) eg New York Kind of place geographical or architectural (G3) eg Forest Place symbolised (A3) eg Paradise
When? Linear time date or period (S4) eg 2010 Cyclical time season or time of day (G4) eg Spring Emotion/abstraction symbolised by time (A4) eg Father Time
11

• Small scale study using 250 images and associated
  tags on Flickr
• Tags categorised using facets from Shatfords
  matrix
• Comparisons made with results of previous
  research into user queries

12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15

• Limited sample size only 250 images
• Use of Flickr as domain for study only 38 of
  users apply tags
• Subjectivity of categorising tags only one
  person assigning tags to categories
• Suitability of Shatfords matrix 22 of terms
  could not be categorised
• Lack of online query studies with which to
  compare the results

16

• Broad similarities between the image facets used
  in queries and image tags
• But differences in the level of specificity
• Need to develop systems to bridge this gap
• Consider the value of tags for browsing systems