AFFECT-HRI: A Comprehensive Human-Robot Interaction Dataset
The article identifies a critical gap in current affective computing research, where existing datasets are often unsuitable for practical robotic applications due to a lack of real-world dynamics and scarcity of open data. Specifically, the absence of publicly available datasets containing labeled physiological metrics hinders the development of robust emotion recognition systems for human-robot interaction. This limitation prevents researchers from accurately modeling human affect in complex, realistic scenarios, thereby stalling progress in creating responsive and empathetic robotic agents. To address this challenge, the study introduces the AFFECT-HRI dataset, the first comprehensive open resource featuring physiological data labeled with human affect in an HRI context. By combining subjective assessments with objective sensor data from a realistic retail scenario, this resource provides the necessary ground truth for training and validating emotion recognition algorithms. The inclusion of diverse conditions, such as moral and liability scenarios, allows for a nuanced understanding of how different interaction types influence user emotions, offering a richer foundation for interdisciplinary research in psychology, law, and computer science. This contribution is highly relevant to open data initiatives because it democratizes access to high-quality, multimodal data that was previously confined to proprietary or inaccessible research groups. By making these labeled physiological signals publicly available, the dataset empowers the broader scientific community to advance affective computing technologies without the prohibitive costs of data collection. This openness accelerates innovation, ensuring that future emotion recognition models are built on diverse, representative, and transparent data sources, ultimately leading to more ethical and effective human-robot interactions.
Source: azorobotics.comPublished on 2024-04-13
Related news
- Can synthetic data boost fairness in medical imaging AI?
- OSCE Trains Ukraine's Civil Society, Private Sector in Open Data Transparency
- ‘We need the transparency’: New bill would strengthen Alabama’s public records law
- Research Urges Clear Guidelines for Fair Synthetic Data Use
- Several Top Museums in UK Admit Hundreds of Items Were Lost, Stolen or Destroyed