DAE-NG: A Shareable and Open Document Image Annotation Data Framework
Bart Lamiroy · 2017
A few years ago we introduced the Document Analysis and Exploitation (DAE) Platform with the goal of providing a comprehensive environment for the hosting and distribution of reproducible data, annotations and software. DAE was conceived and built around a core data model that establishes an exhaustive range of relations between document images, annotation areas, interpretations or Ground Truth, but also links the data to user interactions, experimental protocols or program executions. The resulting platform provided a set of services (querying, up- and download, remote execution...) that allowed users to use as a reference repository for their experimental setups and validations. This paper reflects on several years of existence, usages and non-usages of the DAE platform and presents a new architecture, resulting from a critical analysis of the first version, with the intention of getting it more broadly adopted by the Document Image Analysis community.