Loading...

Emotion Recognition in an Indirect Way

Emadi, Fatemeh | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57303 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Bagheri Shouraki, Saeed
  7. Abstract:
  8. In the rapidly evolving landscape of digital communications and beyond, understanding human emotions has become a critical element for various applications. Although numerous studies and research have been conducted in this area, current methods face multiple challenges. Among these challenges are the heavy computational load and the loss of individual privacy. This study focuses on designing and testing a new method for emotion recognition aimed at addressing these challenges. This new method is based on Daniel Kahneman's theory of fast and slow thinking and an idea proposed by Mr. Heratizadeh, which is grounded in the fast decision-making system of individuals. It operates by using the selection of media data to recognize emotional states. For example, this method can detect the driver's emotions through the music they unconsciously choose while driving. This eliminates the need to process facial signals, voice, or other direct driver signals, thereby better preserving privacy and reducing computational load. In this study, the proposed method was tested over three different periods. The results indicated that while this method addresses many challenges present in previous methods, it is highly sensitive to the input data. For instance, music with different languages, cultures, and histories does not perform well with this method. Even within a common language and culture, individuals might receive different emotions from the same data. This sensitivity to data highlights the importance of the data in the effectiveness of this method
  9. Keywords:
  10. Emotion Detection ; Affective Computing ; Artificial Intelligence ; Valence-Arousal Model

 Digital Object List

 Bookmark

...see more