Table 8.
Comparison with glasses-based eating and chewing monitoring systems.
| Study | Goal | Sensors | Participants, n | Setup | Performance |
| Bedri et al [13], 2020 | Eating and drinking | IMUa, proximity, and camera | 18 (laboratory) and 5 (real-world) | Laboratory and real world | F1-score: 0.89 |
| Shin et al [14], 2022 | Eating | Piezoelectric and IMU | 30 | Real-world | F1-score: 0.92 |
| Bello et al [15], 2023 | Facial expressions and eating and drinking | IMU, pressure, microphone, force, and piezoelectric | 10 | Real-world | F1-score: 0.86 (expressions); 0.94 (eating or drinking) |
| Farooq and Sazonov [16], 2016 | Eating and physical activity | Piezoelectric and IMU | 10 | Laboratory | F1-score: 99% (eating vs activity) |
| Chung et al [17], 2017 | Chewing and 5 other activities | Load cells | 10 | Laboratory | F1-score: 94% |
| Zhang and Amft [18], 2018 | Chewing and eating | EMGb sensors | 10 | Laboratory and real-world | Precision or recall: 95% (laboratory) and 78% (real-world) |
| This study | Chewing and eating | OCO optical sensors | 128 (laboratory) and 8 (real-world) | Laboratory and real-world | F1-score: 0.91 (laboratory) and 0.88 (real-world) |
aIMU: inertial measurement unit.
bEMG: electromyography.