Overview

The MetaDance component of the project focuses on developing a solution to documenting dance by developing a dance annotation tool that implements a dance ontology to extract the terminology from it. By doing so, a dance vocabulary is extracted and is used for annotating dance media files within the tool. These annotations can be further used for quick retrieval of dance media file.

Objectives

Develop a mobile dance application that documents Latin dance based on user annotations and allow users to search dance content.

Mobile Dance Tool

Develop tool within a Mobile platform

Dance Vocabulary

Develop a model in which dance terms can be stored and used as predefined dance terms for annotating

Media file retrieval

Retrieve media files based on user annotations

Ontology Design

The purpose of the ontology is to provide the mobile annotation tool with a dance vocabulary
   
Dance ontology to be designed for two Latin dances only, in order to demonstrate the resuability of the knowledge acquisition process and extensibility of the dance ontology.

Ontology Development

As mentioned above, the purpose of the ontology is to provide a dance vocabulary for the annotation process within the mobile tool. In order to develop the ontology we need to acquire the necessary dance terminology for both Salsa and Cha-Cha-Cha. 
The following steps were followed within the ontology development process:

Ontology Developed

The dance terminology extracted from online resources were then used to develop the lightweight dance ontology.

1099

Axioms

267

Classes

3

Object Properties

The above diagram illustrates the general structure that was developed of the dance ontology. Each Latin dance type will have corresponding difficulty levels. In turn, each difficult level will have dance moves categorized under them, while each dance move is made up of actions.

Download Ontology

Dance Term 
Collection Results

The results to test how well constructed the process of extracting terms were divided into two, i.e. Salsa and Cha-Cha-Cha results. Below are the graphs illustrating the number of new dance terms being extracted at each iteration.

Salsa Term Collection

Salsa included eight iterations of term extraction. As can be seen from the graph on the left, there is a downward logarithmic trend in the number of new terms collected at each iteration of term extraction (orange line). A logarithmic trendline was chosen as it best demonstrates the rate of change of data, and whether it increases, decreases or levels out. 
The R-squared value of the Salsa term collection is 0,9075, which is classified as a large positive association. This means that there tends to be a large increase in the number of terms collected at each point.

Cha-Cha-Cha 
Term Collection

Cha-Cha-Cha included four iterations of term extraction. As can be seen from the graph on the right, there is a downward logarithmic trend in the number of new terms collected at each iteration (orange line). 
the R-squared value of the Cha-Cha-Cha term collection is 0,6494, which is classified as a small positive association. This indicates that there is a slower increase in the number of new terms collected.

Ontology Implementation

In order for the ontology to be used within the system, an Ontology class was created to extract the relevant information from the OWL file. This class will be able to extract dance terms and actions and organize them based on the dance type to which it needs to be categorized under, as well as the difficulty levels. 

MetaDance: 
Dance Annotation Tool

The annotation tool implements the dance ontology developed by extracting the terminology for use within the annotation process.

Functionalities

Annotate Dance Media Files
   
Media Retrieval
   
Edit/Delete Annotations

learn more

 Features

Created with Sketch.  Simplicity

Tool ensures that the tool is simple, and thus easy to use and navigate to complete desires tasks.

 Android OS

Tool operates on Android operating systems, with a minimum API level 14.

 User Feedback

Tool provides users with feedback important tasks have been completed to ensure users that tasks have been completed.

 Documents

Literature Review

Review of the academic literature relevant to dance annotation tools and their implementations.

Read More 

Project Proposal

Assess the feasibility of the project and outlines the plans of the project.

Read More 

Final 
Paper

Final Paper containing more details about the project and sections mentioned above, specifically for MetaDance.

Read More 

Project Poster

Poster illustrating the overview of components developed within this project.

Read More 

MetaDance: 
Dance Annotation Tool

Allows users to select videos from their device for annotating and storing annotations in a local SQLite database. Users are able to browse the database to retrieve the videos that they have annotated. To accomodate for mistakes being made during annotating, users are allowed to edit/delete annotations.
The mobile tool allows users to easily access dance media that they have annotated in order to recall and perform dance productions within a social event, educational purposes and many more situations.
To learn more about how the tool works please click the button below to view a video demonstration.

Demo