VTKEL: 30K documents dataset for Visual-Textual-Knowledge Entity Linking
Version 6 2020-09-09, 14:15
Version 5 2019-08-28, 10:12
Version 4 2019-04-10, 11:06
Version 3 2019-04-01, 14:23
Version 2 2019-04-01, 14:18
Version 1 2019-04-01, 13:36
dataset
posted on 2020-09-09, 14:15 authored by Shahi DostShahi Dost, Luciano Serafini, Marco Rospocher, Lamberto Ballan, Alessandro Sperduti<h3>(Updated version, after fixed some bugs)</h3><h3>VTKL dataset, contains documents composed of pictures with five corresponding textual captions for each image. The VTKL dataset is obtained by extending the Flikr30k dataset, designed for visual-textual mention alignment, with links to YAGO ontolgy, one of the largest <em>web knowledge base</em>. These links are obtained automatically by processing each image caption with PIKES, an <em>NLP</em> tool for <em>entity recognition</em> and <em>linking</em>.<br></h3>
Funding
Fondazione Bruno Kessler, Italy
History
Usage metrics
Categories
Keywords
Knowledge Representation,Computer vision; Image processing; Manufacturing systems; Defect detection; Hot rolling; Rail; ProfileNatural Language Processing, Word EmbeddingsEntity Recognitionentity linkingMultimedia ProgrammingNatural Language ProcessingArtificial Intelligence and Image ProcessingComputer Vision
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC