Related Experiment Video
Updated: Dec 5, 2025
05:25
Author Spotlight: Assessing Brain Activity in Robotic-Assisted Lower Limb Rehabilitation Using fNIRS
Published on: June 7, 2024
1.6K
Introducing the NEMO-Lowlands iconic gesture dataset, collected through a gameful human-robot interaction
Jan de Wit1, Emiel Krahmer2, Paul Vogt3
1Department of Communication and Cognition, Tilburg Center for Cognition and Communication, Tilburg University, PO Box 90153, 5000LE, Tilburg, Netherlands. j.m.s.dewit@tilburguniversity.edu.
Behavior Research Methods
|October 20, 2020
Summary
Researchers created a new dataset of silent iconic gestures using a robot-led charades game. This method captures naturalistic human gesturing and aids robot gesture recognition research.
Area of Science:
- Human-Computer Interaction
- Robotics
- Cognitive Science
Background:
- Silent iconic gestures are crucial for human communication and human-robot interaction.
- Existing gesture datasets often lack naturalistic elicitation methods and diverse participant pools.
Purpose of the Study:
- To introduce a novel dataset of silent iconic gestures.
- To present a robot-based elicitation method for collecting gesture data.
- To enable future research in gesture recognition and production.
Main Methods:
- A charades game was played with a humanoid robot at public venues (science museum, music festival).
- 428 participants (adults and children) performed 3715 silent iconic gestures for 35 objects.
- A novel interactive elicitation method allowed for participant choice in gesture representation and included repair strategies.
Main Results:
- A large, naturalistic dataset of silent iconic gestures was successfully collected.
- The robot-based elicitation method proved effective in diverse, public settings.
- The dataset includes insights into gesture repair strategies when recognition fails.
Conclusions:
- The developed dataset and elicitation method offer valuable resources for human gesturing behavior research.
- The findings support the use of robots in data collection for gesture recognition and production systems.
- The publicly available method facilitates consistent data collection across different contexts and cultures.
