Multitask Siamese Network for Remote Photoplethysmography and Respiration Estimation.

Heart and respiration rates represent important vital signs for the assessment of a person's health condition. To estimate these vital signs accurately, we propose a multitask Siamese network model (MTS) that combines the advantages of the Siamese network and the multitask learning architecture. The MTS model was trained by the images of the cheek including nose and mouth and forehead areas while sharing the same parameters between the Siamese networks, in order to extract the features about the heart and respiratory information. The proposed model was constructed with a small number of parameters and was able to yield a high vital-sign-prediction accuracy, comparable to that obtained from the single-task learning model; furthermore, the proposed model outperformed the conventional multitask learning model. As a result, we can simultaneously predict the heart and respiratory signals with the MTS model, while the number of parameters was reduced by 16 times with the mean average errors of heart and respiration rates being 2.84 and 4.21. Owing to its light weight, it would be advantageous to implement the vital-sign-monitoring model in an edge device such as a mobile phone or small-sized portable devices.

The balance recovery mechanisms against unexpected forward perturbation.

Falls are one of the main concerns of the elderly. Proper postural adjustments to maintain balance involve the activation of appropriate muscles to produce force and to relocate the center of body mass (CoM). In this study, biomechanical aspects of dynamic postural responses against forward perturbations were experimentally determined by simultaneous measurements of joint angles and EMG activations. Thirteen young and healthy volunteers took turns standing on a flat platform, and were directed to move in the forward direction by an AC servo-motor set at two different speeds (0.1 and 0.2 m/s). Joint motions were recorded, and they followed the sequence of ankle dorsiflexion, knee flexion, and then hip flexion during the later acceleration phase (AP) in order to maintain postural balance against forward perturbation. Tibialis anterior for the ankle dorsiflexion and biceps femoris for the knee flexion were activated during the second half of the AP as the primary muscles to recover balance. In addition, gastrocnemius, which was related to ankle plantarflexion, and rectus femoris, which was related to knee extension, were activated to maintain balance. Movements of the center of plantar pressure and ground reaction forces in fast-speed perturbation were significantly larger than those in slow-speed perturbation. As a result, the ankle strategy was used for slow-speed perturbation, but the mixed strategy consisting of both ankles and hip were used for fast-speed perturbation.