About the M3D Training Program

Structure of the Training Program 

In the preparatory section, we will provide the prior knowledge you are going to need to successfully complete the challenges. 

Learning about M3D: To start learning about how to use the M3D labeling system, it is important to first get to know more about what M3D is, what the main approach is and what gestural dimensions it consists of. 

Getting started with ELAN: for annotating gestures with the M3D labeling system, we need a software to actually annotate. We work with ELAN that's why it is important to know how ELAN works and how to use it. 

Assessing Annotations: Annotating can often be a theory-driven and subjective affair. This is why we recommend reading this note before starting to annotate. 

The training program is organized into challenges, or sections that represent the key concepts within the different gesture dimensions set out in M3D. The various challenges are presented in the order of the proposed annotation workflow*. In other words, each step in the workflow is described as a challenge. However, if you are interested in one specific aspect of gesture, feel free to jump right into the corresponding challenge. They are organized in a way that they can be generally understood without having to complete the one that comes before.

Identifying Gesture Units: In this challenge, you will learn about how gestures are grouped to larger Gesture Units and how to identify and annotate those units. 

Identifying Gesture Form: In this challenge, it is all about kinematics. You will learn how to identify the handshape, trajectory shape and direction and handedness of a gesture. 

Identifying Gesture Phases: In this challenge, you will learn how to identify the different phases of a gestural movement and how to annotate them. You will also recognise that only one phase is obligatory to be present because it carries the meaning. 

Identifying Rhythmic Properties - Beat-like-ness: In this challenge, you will find out what Beat-like-ness is and how to evaluate the level of Beat-like-ness of a gesture.

Identifying Referentiality: Talking about meaning, in this challenge, you will learn how to identify the semantic meaning of a gesture. 

Identifying the Pragmatic Domain: In this challenge, you will learn how to identify the pragmatic meaning of a gesture. 

*Proposed annotation workflow in ELAN: Gesture Units > Gesture Form > Gesture Phases > Rhythmic Properties > Gesture Referentiality > Gesture Pragmatics

Within all challenges, the structure is identical:

Explanatory Video

The first part of the challenge is to watch the explanatory video. We talk about the particular gesture dimension, its theoretical underpinnings, and its specific properties. 

Task 1: Recognize the [...] 

For the first task, the idea is to identify the different aspects of the dimension in a series of short videos. The task is organized in multiple-choice questions, with an increasing level of difficulty, where you will get immediate feedback after answering each question. This task is indicated by the blue M3D logo. 

Task 2: Select the best [...] annotation

In the second task, the idea is to look at 3 different annotations and to try to match the best annotation with the video. It helps you practice assessing actual annotations in ELAN. Here, you will also get immediate feedback after answering each question. This task is indicated by the purple M3D logo. 

Annotation Tutorial Video: Let’s annotate together!

In this tutorial video, you will see a annotation demonstration with commentary. This will help you learn how to annotate the particular gesture dimension in ELAN,  and highlights things to consider while annotating each dimension. 

Annotating on your own: Now it’s your turn!

This is the final task of each challenge. The idea is to download a video we provide and to annotate on your own. Once you have completed your annotations, compare your annotations with a sample solution.