This project aimed to develop a human-machine interface based on computer vision to detect facial gestures, facilitating the assistance, care, and monitoring of individuals with severe motor dependence. The system was designed to empower users by enabling them to control their environment, such as household devices (e.g., lights, televisions), and request assistance when necessary.
The project followed a comprehensive, phased approach that included the development and validation of facial recognition software capable of detecting subtle facial movements. Computer vision algorithms were combined with deep learning models to accurately identify key facial gestures such as eyebrow raises and mouth movements. Participants in the validation process included men and women with severe motor impairments, who were tested in both controlled laboratory settings and real-world environments. The system was assessed for its accuracy, usability, and response time, using metrics such as sensitivity, precision, and the System Usability Scale.
The interface was validated with a group of 11 participants with severe motor impairments, including individuals with conditions such as tetraplegia and traumatic brain injury. The system demonstrated robust performance, achieving 100% accuracy in recognizing facial gestures under diverse conditions, including variations in lighting and background. Response times were consistently under one second, enabling real-time interaction between the users and their environment. Participants were able to autonomously control devices such as lights and televisions, as well as request caregiver assistance. Usability evaluations, based on the System Usability Scale, yielded high satisfaction scores, with the majority of participants rating the system above 90. However, areas for improvement were identified, particularly regarding the customization of gesture commands for users with more limited facial mobility.
The project successfully developed a human-machine interface that significantly enhances the autonomy of individuals with severe motor impairments. By allowing users to interact with their environment without caregiver assistance, the system promotes independence, reduces caregiver burden, and improves the quality of life. The technology demonstrated high reliability and satisfaction rates among participants, making it a viable solution for both home and clinical settings. However, further customization is needed to better accommodate users with more restricted facial mobility.
This research has significant implications for physiotherapy practice and healthcare management. The implementation of this system in clinical and home care settings could increase care efficiency and enhance the autonomy of elderly individuals, contributing to better mental health outcomes. Future work should focus on expanding the range of recognized gestures, increasing personalization for individual users, and conducting long-term studies to assess the system's impact on the mental and physical well-being of users.
Motor Dependence
Computer Vision
