DefiningImageAccess/Project/eChase
From ImageWeb
| DefiningImageAccess/Project/eChase | |
|---|---|
| homepage:=http://www.echase.org/}} | |
| [[has sub-project::{{{Subproject}}}]]}} | |
| sub-project of::DefiningImageAccess/Project/SCULPTEUR}} | |
| [[start date:={{{Start date}}}]]}} | |
| [[end date:={{{End date}}}]]}} | |
| Status:=In progress}} | |
| JISCProject:=false}} | |
| [[Image Materials:={{{ImageMaterials}}}]]}} | |
| Focus:=Metadata Creation, Metadata Harvesting, Metadata Querying and Virtual Exhibition}} | |
| [[Publishes::{{{Publishes}}}]]}} | |
| [[References::{{{References}}}]]}} | |
| [[Uses::{{{Uses}}}]]}} | |
| [[Creates::{{{Creates}}}]]}} | |
| [[Partner::{{{Partner}}}]]}} | |
| [[Contact::{{{Contact}}}]]}} | |
| Defining Image Access | DefiningImageAccess/RelatedWork |
eChase
DefiningImageAccess/Project/eChase
A continuing project of the SCULPTEUR project.
- http://www.echase.org/
- http://www.echase.org/documents.html - this links to an annual report
- http://eprints.ecs.soton.ac.uk/11567/01/echase.pdf
- http://piltdownman.it-innovation.soton.ac.uk/echase/
- http://www.echase.org/03_description.html
- http://www.echase.org/pdf/eCHASE_AnnualReport_05.pdf
See also:
The goal of eChase is very much close to what SCULPTEUR wants to achieve.
[1] The document about the progress of the eChase project in its first phase. http://www.echase.org/documents.html
The source code from the SCULPTEUR project, and the precedent Artiste project, is shared at SourceForge:
Conclusions from the project final report
A copy of the project "Systems Integration and Service Operation Report" was provided to us privately, and does not appear to be available on the web or other publiucation medium.
The following are some conclusions liftedn from that report:
- "The metadata mapping process is to first convert data sets into RDF using ontologies such as CIDOC CRM, CIDOC CRM Core and SKOS. The resulting data in RDF must then be processed further using a workflow-based approach to overcome irregularities between the different collections’ approaches to encode data such as dates or places."
- "There is a clear benefit in knowing the thesaurus structure to help understand the keyword entries. There are also lots of repeated keywords [...] keywords such as "Wood" will actually appear several times in the thesaurus:
- Applied Arts and Crafts > Object > Wood
- Botany > Vegetation > Element or Composition Related to Vegetable Kingdom > Wood
- Industry > Structure and Scene > Wood
- Vegetation > Vegetal Element or Composition > Wood
- We decided that the thesaurus structure was crucial for the eCHASE system. We were unable to obtain this information from [the content provider]. Instead we investigated techniques to automatically reconstruct the thesaurus structure from the keyword list for each image, which was possible as the keyword lists had been exported in a uniform manner. However, due to the duplicated keywords the results were not perfect."
- In some cases, useful metadata could be obtained only by implementing custom scrapers for web page content. "By scraping this web page, using the Ruby Hpricot library (http://code.whytheluckystiff.net/hpricot/), we were able to extract a category for each image, the photographer’s name, the date the photograph was taken and the location. The location name could then be translated into English and looked up in the Gazetteer as before." Various other ad hoc techniques were applied.
- "[another content provider] provided their data as an Access database dump, which was very well structured and provided very few issues for importing into eCHASE. Descriptions and keywords were provided in a number of languages, although the coverage of the translations was not complete."
- "During the project we were able to test the metadata import modules with other collections. We were interested to see how well the metadata structure, based on CIDOC CRM, was able to model new collections, and how the mapping and processing tools coped with new data. It was our view that the experience gained in mapping each new metadata collection would make the next collection easier and faster to import."
- on importing a brand new collection: "The content was provided as an XML dump, which was easily converted into RDF using a custom-written XSLT. Once the data was converted into RDF, it was possible to apply the Taverna workflows composed for the eCHASE collections (e.g. to normalise dates and places) on the [...] data, and import the content into the eCHASE system.
- "The development of the eCHASE system was carried out using an iterative approach that focused on providing prototypes to the users, receiving feedback, and converting those comments into requirements for the future prototypes. This style of development ensured that the system design was grounded firmly in the user's needs." [...] "This change in approach has significantly improved the way that the technical work in eCHASE is driven by user need and the results have been hugely beneficial with the end user partners reporting that the eCHASE software now provides the functionality needed for the business trials and is generally of a very high standard." (These quotes are taken from slightly different contexts in the source report, but I believe the effect is consistent with its intended meaning.)
Focus:=Metadata Creation Focus:=Metadata Harvesting Focus:=Metadata Query Focus:=Virtual Exhibition
