Skip to content
No description, website, or topics provided.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
config
img
src
.classpath
.project
README.md

README.md

MULTISENSOR visual-based ontology alignment and GUI

The MULTISENSOR visual-based ontology alignment implements the ontology alignment algorithm for computing a visual-based similarity metric for entity matching between two ontologies. Each ontological entity is associated with sets of images, retrieved through ImageNet or web-based search, and visual feature extraction, clustering and indexing for computing the similarity between concepts is employed. An adaptation of a popularWordnet-based matching algorithm to exploit the visual similarity has also been developed. More details about this algorithm can be found in [1]. A GUI for the Alignment API is also implemented.

[1] C. Doulaverakis, S. Vrochidis, I. Kompatsiaris, "Exploiting visual similarities for ontology alignment", 7th International Conference on Knowledge Engineering and Ontology Development (KEOD 2015), Lisbon, Portugal, 12-14 November, 2015

#Description

The implementation provided is based on the [Alignment API v4.6] (http://alignapi.gforge.inria.fr/). An extension to Alignment API is implemented which allows to combine different ontology matching algorithms using a weighted sum approach where each matcher's score is multiplied before the overall sum is computed. Weights can be set through the method setMatcherWeights of the class gr.iti.multisensor.matrix.MSWeighting. The Alignment API is not a Maven project so to build it follow the instruction [here] (http://alignapi.gforge.inria.fr/maven.html)

For the Visual alignment algorithm, you will have to import the project [multimedia-indexing] (https://github.com/MKLab-ITI/multimedia-indexing) for computing feature extraction and indexing. For retrieving the images that correspond to the ontological entities, [ImageNet] (http://www.image-net.org/) and Yahoo Image Search are accessed. For Yahoo Image Search you will have to obtain a BOSS account and provide the Key and Secret in the class gr.iti.multisensor.ui.utils.Parameters. You will have to download [Wordnet] (https://wordnet.princeton.edu/) and the [Stanford POS tagger] (http://nlp.stanford.edu/software/tagger.shtml). For Wordnet, set the appropriate Wordnet dir path in the file config/config.txt and in JWNL_properties.xml.

For running the GUI, the main class is gr.iti.multisensor.ui.MultiAlignMainWindow

Version

1.0.0

You can’t perform that action at this time.