A spatial relation-based framework to perform visual information extraction

作者:Giuseppe Della Penna, Daniele Magazzeni, Sergio Orefice

摘要

The Spatial Relation Query (SRQ) tool is a graphical software environment, supported by a SQL-like language, which enables users to perform information extraction driven by the visual appearance and the spatial arrangement of the information. The tool has been initially customised to work on specific application domains, like web pages and geospatial data. In this paper, we present the theoretical formalisation of the visual information extraction (VIE) task and accordingly the redesign of the SRQ tool, which is now a full-featured, general-purpose information extraction system. Moreover, we show a new application of the VIE framework to the analysis and visual information extraction from PDF files.

论文关键词:Information extraction, Spatial relations, Visual information retrieval, PDF analysis

论文评审过程:

论文官网地址:https://doi.org/10.1007/s10115-011-0394-4