CA Manager

Content Annotation Manager uses ontology-based semantic annotation to transform heterogeneous content (text, image, video…) into integrated and organized content. While numerous tools exist to annotate content and transcode data, they fail to provide a transcodification workflow, the flexibility to customize annotation, and the ability to coordinate multiple semantic technologies and resources. Moreover, many systems lack compliance with the open Semantic Web standards.

CA Manager is a framework for building and managing customized workflows for semantic annotation of content. A typical workflow includes the following stages:

Coordinate a transcription of existing annotations (metadata, structured information) into a normalized reference system

Use text mining to extract information out of unstructured text

 Access a domain ontology to infer new annotations

Automatically check completeness of the information, and submit the annotation result for validation by a human expert if the quality control fails

Transform final content annotation into the RDF format, ready to be added to content and stored in a metadata repository

For image annotation, the process uses image analysis for extraction and classification of image zones, as well as the existing image metadata (time, place…) and a domain ontology to infer new metadata.

In addition to content annotation, CA Manager supports the ontology enrichment process by proposing new terms and facts to update and enrich terminologies, taxonomies and the knowledge base. This feature uses a third-party text-mining engine in conjunction with validation by human experts.

Tag content using linguistic extraction tools and/orLinked Data on the web

Structure information  with annotations controlled by linguistic extraction tools

Control annotations from linguistic extraction against ontology, normalize, construct knowledge graph from annotations, add new knowledge to knowledge store

Validate (optional) - review annotations / knowledge, approve updates to knowledge base

Top advantages

  •  UIMA-based framework interconnects various information extraction tools, ontologies, reasoning engines, internal and external resources, and human expertise
  •  The extensibility of the UIMA-based framework makes it possible to add new connectors to external tools
  •  Adapted for Open Data and Semantic Web projects
  •  Mediation between semantic annotation and ontology population
  •  Control and consolidation algorithms to avoid loss of information and check annotation quality and completeness
  •  Annotation UI which allows humans to be involved in the validation step

> Feature summary