Abstract
Interactive games have the potential to mitigate or prevent gait impairments and cognitive decline in older adults. This study aimed at developing a novel real-time step detection and direction classification approach to be used in the evaluation of multidirectional steps and interaction while playing motor-cognitive games. Two shoe-mounted inertial sensors were used to capture foot motions, which were treated interchangeably after the application of a novel foot sagittal reflection method. A single multi-class classifier was able to distinguish step direction with an accuracy of 98.1%. Experimental results support the applicability of the solution in the context of interactive motor-cognitive training.

Abstract
Although deep learning is being widely adopted for every topic in pattern recognition, its use for secure and cance-lable biometrics is currently reserved for feature extraction and biometric data preprocessing, limiting achievable performance. In this paper, we propose a novel formulation of the triplet loss methodology, designated as secure triplet loss, that enables biometric template cancelability with end-to-end convolutional neural networks, using easily changeable keys. Trained and evaluated for electrocardiogram-based biometrics, the network revealed easy to optimize using the modified triplet loss and achieved superior performance when compared with the state-of-the-art (10.63% equal error rate with data from 918 subjects of the UofTDB database). Additionally, it ensured biometric template security and effective template cancelability. Although further efforts are needed to avoid template linkability, the proposed secure triplet loss shows promise in template cancelability and non-invertibility for biometric recognition while taking advantage of the full power of convolutional neural networks.

Abstract
Introduction: Recently, kinematic analysis of the drinking task (DRINK) has been recommended to assess the quality of upper limb (UL) movement after stroke, but the accomplishment of this task may become difficult for poststroke patients with hand impairment. Therefore, it is necessary to study ADLs that involve a simpler interaction with a daily life target, such as the turning on a light task (LIGHT). As the knowledge of movement performed by healthy adults becomes essential to assess the quality of movement of poststroke patients, the main goal of this article was to compare the kinematic strategies used by healthy adults in LIGHT with those that are used in DRINK. Methods: 63 adults, aged 30 to 69 years old, drank water and turned on a light, using both ULs separately, while seated. The movements of both tasks were captured by a 3D motion capture system. End-point and joint kinematics of reaching and returning phases were analysed. A multifactorial analysis of variance with repeated measures was applied to the kinematic metrics, using age, sex, body mass index and dominance as main factors. Results: Mean and peak velocities, index of curvature, shoulder flexion and elbow extension were lower in LIGHT, which suggests that the real hand trajectory was smaller in this task. In LIGHT, reaching was less smooth and returning was smoother than DRINK. The instant of peak velocity was similar in both tasks. There was a minimal anterior trunk displacement in LIGHT, and a greater anterior trunk displacement in DRINK. Age and sex were the main factors which exerted effect on some of the kinematics, especially in LIGHT. Conclusion: The different target formats and hand contact in DRINK and LIGHT seem to be responsible for differences in velocity profile, efficiency, smoothness, joint angles and trunk displacement. Results suggest that the real hand trajectory was smaller in LIGHT and that interaction with the switch seems to be less demanding than with the glass. Accordingly, LIGHT could be a good option for the assessment of poststroke patients without grasping ability. Age and sex seem to be the main factors to be considered in future studies for a better match between healthy and poststroke adults.

Abstract

Abstract

Abstract
This study presents subject specific analysis at complementary time, frequency and phase plane domains of entire lower limb joint angular kinematic series during normal and modified gait. Healthy adult subject case study was assessed at human movement lab during normal gait, stiff knee gait and slow running. In vivo and non-invasive assessment was performed with eight camera Qualisys system at 100 Hz acquisition of cartesian coordinates of skin reflective markers at lower limb anatomical selected point. Inverse kinematics was performed using AnyGait v.0.92 and TLEM model to match the size and joint morphology of the stick-figure. Entire series of subject specific hip, knee and ankle joint angular displacements (?), angular velocities (?) and angular accelerations (a) at the sagittal plane were assessed at the time, frequency and phase plane domains providing useful information for development of subject specific rehabilitation equipment towards gait restoring. The main novelty and contribution of this study when compared with available literature consists on complementary approach of continuous entire signal analysis, time, frequency and phase analysis which can contribute for prescription of requisites for design of rehabilitation equipment towards normal gait of subjects with stiff knee gait and slow running. © 2019, CISM International Centre for Mechanical Sciences.