Semi-automated annotation of page-based documents within the Genre and Multimodality framework

Abstract

This paper describes ongoing work on a tool developed for annotating document images for their multimodal features and compiling this information into a corpus. The tool leverages open source computer vision and natural language processing libraries to describe the content and structure of multimodal documents and to generate multiple layers of XML annotation. The paper introduces the annotation schema, describes the document processing pipeline and concludes with a brief description of future work.