Journal article Open Access

Affective Robots: Evaluation of Automatic Emotion Recognition Approaches on a Humanoid Robot towards Emotionally Intelligent Machines

Silvia Santano Guillén; Luigi Lo Iacono; Christian Meder


Citation Style Language JSON Export

{
  "publisher": "Zenodo", 
  "DOI": "10.5281/zenodo.1316752", 
  "container_title": "International Journal of Mechanical, Industrial and Aerospace Sciences", 
  "language": "eng", 
  "title": "Affective Robots: Evaluation of Automatic Emotion Recognition Approaches on a Humanoid Robot towards Emotionally Intelligent Machines", 
  "issued": {
    "date-parts": [
      [
        2018, 
        4, 
        4
      ]
    ]
  }, 
  "abstract": "One of the main aims of current social robotic research<br>\nis to improve the robots&rsquo; abilities to interact with humans. In order<br>\nto achieve an interaction similar to that among humans, robots<br>\nshould be able to communicate in an intuitive and natural way<br>\nand appropriately interpret human affects during social interactions.<br>\nSimilarly to how humans are able to recognize emotions in other<br>\nhumans, machines are capable of extracting information from the<br>\nvarious ways humans convey emotions&mdash;including facial expression,<br>\nspeech, gesture or text&mdash;and using this information for improved<br>\nhuman computer interaction. This can be described as Affective<br>\nComputing, an interdisciplinary field that expands into otherwise<br>\nunrelated fields like psychology and cognitive science and involves<br>\nthe research and development of systems that can recognize and<br>\ninterpret human affects. To leverage these emotional capabilities<br>\nby embedding them in humanoid robots is the foundation of<br>\nthe concept Affective Robots, which has the objective of making<br>\nrobots capable of sensing the user&rsquo;s current mood and personality<br>\ntraits and adapt their behavior in the most appropriate manner<br>\nbased on that. In this paper, the emotion recognition capabilities<br>\nof the humanoid robot Pepper are experimentally explored, based<br>\non the facial expressions for the so-called basic emotions, as<br>\nwell as how it performs in contrast to other state-of-the-art<br>\napproaches with both expression databases compiled in academic<br>\nenvironments and real subjects showing posed expressions as well<br>\nas spontaneous emotional reactions. The experiments&rsquo; results show<br>\nthat the detection accuracy amongst the evaluated approaches differs<br>\nsubstantially. The introduced experiments offer a general structure<br>\nand approach for conducting such experimental evaluations. The<br>\npaper further suggests that the most meaningful results are obtained<br>\nby conducting experiments with real subjects expressing the emotions<br>\nas spontaneous reactions.", 
  "author": [
    {
      "family": "Silvia Santano Guill\u00e9n"
    }, 
    {
      "family": "Luigi Lo Iacono"
    }, 
    {
      "family": "Christian Meder"
    }
  ], 
  "volume": "11.0", 
  "version": "10009027", 
  "type": "article-journal", 
  "issue": "6", 
  "id": "1316752"
}
172
105
views
downloads
All versions This version
Views 172173
Downloads 105105
Data volume 20.1 MB20.1 MB
Unique views 161162
Unique downloads 9999

Share

Cite as