View
217
Download
0
Tags:
Embed Size (px)
Citation preview
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
The ARTISTE ProjectThe ARTISTE Project
Building a system for
Art Image Storage
Retrieval
Analysis
and Navigation
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
The ConsortiumThe Consortium
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
The ObjectivesThe Objectives
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
The SystemThe System
• Will be a distributed database of Art Images and metadata.
• Will have www access.
• Will provide content and meta-data based retrieval navigation and analysis tools.
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
The Project Stage...The Project Stage...
• The project is in its infancy. We are prototyping novel algorithms to meet the
specific needs of the end users.
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
Demonstration of Sub-Image Demonstration of Sub-Image MatchingMatching
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……The ProblemThe Problem
• To find an image (the target) from a collection of images.
• A given image (the query) serves as input, and may be a sub-image of a
larger image.
• The process finds images when the query is not necessarily identical to the
target image.
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……Example 1Example 1
• Query Image
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
...The Result...The Result
• Best matching image with sub-image identified.
NB. Query is before restoration work, target is a restored image. Query and target image also differ in resolution
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……Example 2Example 2
• Query Image
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……The ResultThe Result
• Best match found, with sub-image identified
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……Subsequent Best MatchesSubsequent Best Matches
Retrieved results start from top-left to bottom right.
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
……The AlgorithmThe Algorithm
• This M-CCV technique is being developed in the IAM Group at the University of Southampton, UK.
• It matches colour coherence vectors from a collection of image patches at a range of scales in-order to find the best match.
Project IST_1999_11978 - ARTISTE – An Integrated Art Analysis and Navigation Environment
Review Meeting N.1: Paris, C2RMF, November 28, 2000
Other AlgorithmsOther Algorithms
• …will provide for example: An ability to retrieve images containing
particular textures. An ability to locate and count specific
features of interest to end users. E.g. butterfly supports in the restoration framework.