| Full text | |
| Author(s): |
Preto, Murilo de Souza
;
Ferreira, Fernando Teubl
;
Kurashima, Celso Setsuo
Total Authors: 3
|
| Document type: | Journal article |
| Source: | 25TH SYMPOSIUM ON VIRTUAL AND AUGMENTED REALITY, SVR 2023; v. N/A, p. 5-pg., 2023-01-01. |
| Abstract | |
In recent years, there has been a transition from face-to-face class meetings to videoconferences, which has been further expanded during the COVID-19 pandemic. In the area of education, remote classes entailed new advantages and some challenges - in one hand, allowing for an extended time and location flexibility - while, in the other hand, operating in a inexpressive, distanced, manner. Aiming to increase non-verbal communication in virtual class meetings, this research explores the technical feasibility and potential implementation challenges of developing a virtual class tool for facial expressions telemonitoring through embedded systems. As a suggested solution, a system which allows for a teacher to remotely follow the facial expressions of students through real-time rendered graphical data was developed. Technical experimental results include: an average recognition accuracy of 57%, for all facial expressions, peaking at 80% for happiness, in a median processing time of 2.85 seconds with a Raspberry Pi 4B, and of 0.572 seconds with a commodity computer. These findings highlight the technical viability of the implemented system, possibly allowing to decrease the lack of non-verbal communication in distanced learning. (AU) | |
| FAPESP's process: | 22/10909-5 - Differentiable learning and processing of temporal, spatial, spectral and time-frequency signal representations |
| Grantee: | Miguel Arjona Ramírez |
| Support Opportunities: | Regular Research Grants |