Annotation

 

The CADMOS project aims at designing an annotation schema that accounts for the two-fold contribution of story and characters to narrative analysis. The first step of the project consists in devising a character-based annotation schema and annotating a reference schema of dramatic media objects. Then, the project will test the utility of the annotation production through two pilot works, one original production and one production from archived footage. In order to as- sist the annotation process and to improve the correctness of the annotation, the project develops a web-based software platform to annotate and access the archived objects.

The typical workflow of linear storytelling productions relies upon the informal documents drawn up by the author, namely the characters' bible, which reports the characters’ typical attitudes, personality and behaviours, and the screenplay, organized into scenes. The production process yields the audiovisual assets corresponding to the scenes, through a series of specialized phases (pre-visualization, shooting or computer graphics generation to post-production). Each phase is supported by specific software, but the lack of a formal annotation from the beginning introduces a semantic gap between the final audiovisual segments and their narrative content.

The goal of the CADMOS project is to bridge the semantic gap between the creation of a media object and the role it plays in the story. The project will introduce a high-level, unifying schema for annotating a storytelling artefact. This formal language will rely on the central notion of character, described in terms of long-term goals and related behaviours. The annotation language will rely upon an ontology-based terminological base and the annotation process will be carried out through a web-based software platform. The resulting annotation will be validated by a story engine, which tests the consistency and coherence of characters'
goals and actions within the story advancement.

The analysis perspective will be addressed through the development of a semi-automatic annotation process that will suggest the annotator an effective labelling based on the narrative structure; the annotated corpus will be explored through an interface for the search of segments and proposal of re-use.
The production perspective will take advantage from the design and im- plementation of a novel workflow, derived by the current practices in new media, for carrying out annotation-enriched productions from scratch (producers of premium content), and the design and implementation of an interface for search and retrieval of annotated archive material for carrying out productions based on repertoires (content aggregators). The workflow and the interface will be applied to the production of two pilot projects, respectively.

                                                                      An example of annotation >>