Affordable Access

Access to the full text

Facial emotion recognition through artificial intelligence

Authors
  • Ballesteros, Jesús A.1
  • Ramírez V., Gabriel M.2
  • Moreira, Fernando3
  • Solano, Andrés4
  • Pelaez, Carlos A.4
  • 1 Maestria en Inteligencia Artificial, Universidad Internacional de La Rioja, Logroño , (Spain)
  • 2 Facultad de Ingeniería, Universidad de Medellín, Medellín , (Colombia)
  • 3 REMIT, IJP, Universidade Portucalense, Porto and IEETA, Universidade de Aveiro, Aveiro , (Portugal)
  • 4 Departamento de Operaciones y Sistemas, Universidad Autónoma de Occidente, Cali , (Colombia)
Type
Published Article
Journal
Frontiers in Computer Science
Publisher
Frontiers Media S.A.
Publication Date
Jan 31, 2024
Volume
6
Identifiers
DOI: 10.3389/fcomp.2024.1359471
Source
Frontiers
Keywords
Disciplines
  • Computer Science
  • Original Research
License
Green

Abstract

This paper introduces a study employing artificial intelligence (AI) to utilize computer vision algorithms for detecting human emotions in video content during user interactions with diverse visual stimuli. The research aims to unveil the creation of software capable of emotion detection by leveraging AI algorithms and image processing pipelines to identify users' facial expressions. The process involves assessing users through images and facilitating the implementation of computer vision algorithms aligned with psychological theories defining emotions and their recognizable features. The study demonstrates the feasibility of emotion recognition through convolutional neural networks (CNN) and software development and training based on facial expressions. The results highlight successful emotion identification; however, precision improvement necessitates further training for contexts with more diverse images and additional algorithms to distinguish closely related emotional patterns. The discussion and conclusions emphasize the potential of A.I. and computer vision algorithms in emotion detection, providing insights into software development, ongoing training, and the evolving landscape of emotion recognition technology. Further training is necessary for contexts with more diverse images, alongside additional algorithms that can effectively distinguish between facial expressions depicting closely related emotional patterns, enhancing certainty and accuracy.

Report this publication

Statistics

Seen <100 times