On the image content of the ESA EUSC JRC Workshop on Image Infomation Mining
Abstract
Due to the specific topic the meeting is devoted to, the images appearing in the papers and presentations published through the years on the websites of the ESA EUSC JRC IIM Workshop often play a central role in the ongoing public discourse on methodology and applications. While tools for analyzing textual content are widespread, we introduce instead a methodology for semi–automatically organizing published material based on its image content. The sources of quality degradation that are introduced by the publication process are taken into account and partially inverted. Specific methods are introduced to take into account the Earth observation–related specificities of the published image data. A semi-supervised data analysis and labeling step is used to produce semantically enriched descriptions in a set of linked RDF ontologies. This information is not intended anymore for mere presentation on paper or on screen. It can be aggregated and re–used, e.g. to answer complex queries mixing content– and metadata–based attributes of the items in the dataset.
BIB_text
@Article {
author = {Marco Quartulli and Mikel Zorrilla and Igor García},
title = {On the image content of the ESA EUSC JRC Workshop on Image Infomation Mining},
keywds = {
Remote sensing, Content-Based Image Retrieval
}
abstract = {
Due to the specific topic the meeting is devoted to, the images appearing in the papers and presentations published through the years on the websites of the ESA EUSC JRC IIM Workshop often play a central role in the ongoing public discourse on methodology and applications. While tools for analyzing textual content are widespread, we introduce instead a methodology for semi–automatically organizing published material based on its image content. The sources of quality degradation that are introduced by the publication process are taken into account and partially inverted. Specific methods are introduced to take into account the Earth observation–related specificities of the published image data. A semi-supervised data analysis and labeling step is used to produce semantically enriched descriptions in a set of linked RDF ontologies. This information is not intended anymore for mere presentation on paper or on screen. It can be aggregated and re–used, e.g. to answer complex queries mixing content– and metadata–based attributes of the items in the dataset.
}
date = {2012-10-24},
year = {2012},
}
author = {Marco Quartulli and Mikel Zorrilla and Igor García},
title = {On the image content of the ESA EUSC JRC Workshop on Image Infomation Mining},
keywds = {
Remote sensing, Content-Based Image Retrieval
}
abstract = {
Due to the specific topic the meeting is devoted to, the images appearing in the papers and presentations published through the years on the websites of the ESA EUSC JRC IIM Workshop often play a central role in the ongoing public discourse on methodology and applications. While tools for analyzing textual content are widespread, we introduce instead a methodology for semi–automatically organizing published material based on its image content. The sources of quality degradation that are introduced by the publication process are taken into account and partially inverted. Specific methods are introduced to take into account the Earth observation–related specificities of the published image data. A semi-supervised data analysis and labeling step is used to produce semantically enriched descriptions in a set of linked RDF ontologies. This information is not intended anymore for mere presentation on paper or on screen. It can be aggregated and re–used, e.g. to answer complex queries mixing content– and metadata–based attributes of the items in the dataset.
}
date = {2012-10-24},
year = {2012},
}
