An ensemble-based regression approach for continuous estimation of wrist and fingers movements from surface electromyography.

Late development and evolution of high degree-of-freedom (DOF) robotic hands have seen great technological strides to enhance the quality of life for amputated people. A robust hand kinematic estimation mechanisms have shown promising results to control robotic hands that can mimic the human hand functions and perform daily life hand dexterous tasks. In this paper, we propose an ensemble-based regression approach for continuous estimation of wrist and fingers movements from surface Electromyography (sEMG) signals. The proposed approach extracts time-domain features from the sEMG signals, and uses Gradient Boosted Regression Tree (GBRT) ensembles to estimate the kinematics of the wrist and fingers. Furthermore, we propose two different performance evaluation procedures to demonstrate the efficacy of the approach in providing a feasible approach towards accurately estimating hand kinematics.

A fall prediction methodology for elderly based on a depth camera.

With the aging of society population, efficient tracking of elderly activities of daily living (ADLs) has gained interest. Advancements of assisting computing and sensor technologies have made it possible to support elderly people to perform real-time acquisition and monitoring for emergency and medical care. In an earlier study, we proposed an anatomical-plane-based human activity representation for elderly fall detection, namely, motion-pose geometric descriptor (MPGD). In this paper, we present a prediction framework that utilizes the MPGD to construct an accumulated histograms-based representation of an ongoing human activity. The accumulated histograms of MPGDs are then used to train a set of support-vector-machine classifiers with a probabilistic output to predict fall in an ongoing human activity. Evaluation results of the proposed framework, using real case scenarios, demonstrate the efficacy of the framework in providing a feasible approach towards accurately predicting elderly falls.

Fall detection for elderly using anatomical-plane-based representation.

Falls are a common cause of injuries and traumas for elderly and could be life threatening. Delivering a prompt medical support after a fall is essential to prevent lasting injuries. Therefore, effective fall detection could provide urgent support and dramatically reduce the risk of such mishaps. In this paper, we propose a hierarchical classification framework based on a novel anatomical-plane-based representation for elderly fall detection. The framework obtains human skeletal joints, using Microsoft Kinect sensors, and transforms them to a human representation. The representation is then utilized to classify the sensor input sequences and provide a semantic meaning of different human activities. Evaluation results of the proposed framework, using real case scenarios, demonstrate the efficacy of the framework in providing a feasible approach towards accurately detecting elderly falls.

Study of the usability of spaced retrieval exercise using mobile devices for Alzheimer's disease rehabilitation.

BACKGROUND: Alzheimer's disease (AD) is an irreversible brain disease that slowly destroys memory and thinking skills, and eventually the ability to carry out the simplest daily tasks. Recent studies showed that people with AD might actually benefit from physical exercises and rehabilitation processes. Studies show that rehabilitation would also add value in making the day for an individual with AD a little less foggy, frustrating, isolated, and stressful for as long as possible. OBJECTIVE: The focus of our work was to explore the use of modern mobile technology to enable people with AD to improve their abilities to perform activities of daily living, and hence to promote independence and participation in social activities. Our work also aimed at reducing the burden on caregivers by increasing the AD patients' sense of competence and ability to handle behavior problems. METHODS: We developed ADcope, an integrated app that includes several modules that targeted individuals with AD, using mobile devices. We have developed two different user interfaces: text-based and graphic-based. To evaluate the usability of the app, 10 participants with early stages of AD were asked to run the two user interfaces of the spaced retrieval memory exercise using a tablet mobile device. RESULTS: We selected 10 participants with early stages of AD (average age: 75 years; 6/10, 60% males, 4/10, 40% females). The average elapsed time per question between the text-based task (14.04 seconds) and the graphic-based task (12.89 seconds) was significantly different (P=.047). There was also a significant difference (P<.001) between the average correct answer score between the text-based task (7.60/10) and the graphic-based task (8.30/10), and between the text-based task (31.50/100) and the graphic-based task (27.20/100; P<.001). Correlation analysis for the graphic-based task showed that the average elapsed time per question and the workload score were negatively correlated (-.93, and -.79, respectively) to the participants' performance (P<.001 and P=.006, respectively). CONCLUSIONS: We found that people with early stages of AD used mobile devices successfully without any prior experience in using such devices. Participants' measured workload scores were low and posttask satisfaction in fulfilling the required task was conceivable. Results indicate better performance, less workload, and better response time for the graphic-based task compared with the text-based task.