Gait Activity Classification on Unbalanced Data from Inertial Sensors Using Shallow and Deep Learning.

Activity recognition is one of the most active areas of research in ubiquitous computing. In particular, gait activity recognition is useful to identify various risk factors in people's health that are directly related to their physical activity. One of the issues in activity recognition, and gait in particular, is that often datasets are unbalanced (i.e., the distribution of classes is not uniform), and due to this disparity, the models tend to categorize into the class with more instances. In the present study, two methods for classifying gait activities using accelerometer and gyroscope data from a large-scale public dataset were evaluated and compared. The gait activities in this dataset are: (i) going down an incline, (ii) going up an incline, (iii) walking on level ground, (iv) going down stairs, and (v) going up stairs. The proposed methods are based on conventional (shallow) and deep learning techniques. In addition, data were evaluated from three data treatments: original unbalanced data, sampled data, and augmented data. The latter was based on the generation of synthetic data according to segmented gait data. The best results were obtained with classifiers built with augmented data, with F-measure results of 0.812 (σ = 0.078) for the shallow learning approach, and of 0.927 (σ = 0.033) for the deep learning approach. In addition, the data augmentation strategy proposed to deal with the unbalanced problem resulted in increased classification performance using both techniques.

Monitoring behavioral symptoms of dementia using activity trackers.

Tertiary disease prevention for dementia focuses on improving the quality of life of the patient. The quality of life of people with dementia (PwD) and their caregivers is hampered by the presence of behavioral and psychological symptoms of dementia (BPSD), such as anxiety and depression. Non-pharmacological interventions have proved useful in dealing with these symptoms. However, while most PwD exhibit BPSD, their manifestation (in frequency, intensity and type) varies widely among patients, thus the need to personalize the intervention and its assessment. Traditionally, instruments to measure behavioral symptoms of dementia, such as NPI-NH and CMAI, are used to evaluate these interventions. We propose the use of activity trackers as a complement to monitor behavioral symptoms in dementia research. To illustrate this approach we describe a nine week Cognitive Stimulation Therapy conducted with the assistance of a social robot, in which the ten participants wore an activity tracker. We describe how data gathered from these wearables complements the assessment of traditional behavior assessment instruments with the advantage that this assessment can be conducted continuously and thus be used to tailor the intervention to each PwD.

Semi-Automated Data Labeling for Activity Recognition in Pervasive Healthcare.

Activity recognition, a key component in pervasive healthcare monitoring, relies on classification algorithms that require labeled data of individuals performing the activity of interest to train accurate models. Labeling data can be performed in a lab setting where an individual enacts the activity under controlled conditions. The ubiquity of mobile and wearable sensors allows the collection of large datasets from individuals performing activities in naturalistic conditions. Gathering accurate data labels for activity recognition is typically an expensive and time-consuming process. In this paper we present two novel approaches for semi-automated online data labeling performed by the individual executing the activity of interest. The approaches have been designed to address two of the limitations of self-annotation: (i) The burden on the user performing and annotating the activity, and (ii) the lack of accuracy due to the user labeling the data minutes or hours after the completion of an activity. The first approach is based on the recognition of subtle finger gestures performed in response to a data-labeling query. The second approach focuses on labeling activities that have an auditory manifestation and uses a classifier to have an initial estimation of the activity, and a conversational agent to ask the participant for clarification or for additional data. Both approaches are described, evaluated in controlled experiments to assess their feasibility and their advantages and limitations are discussed. Results show that while both studies have limitations, they achieve 80% to 90% precision.

Variability Analysis of Therapeutic Movements using Wearable Inertial Sensors.

A variability analysis of upper limb therapeutic movements using wearable inertial sensors is presented. Five healthy young adults were asked to perform a set of movements using two sensors placed on the upper arm and forearm. Reference data were obtained from three therapists. The goal of the study is to determine an intra and inter-group difference between a number of given movements performed by young people with respect to the movements of therapists. This effort is directed toward studying other groups characterized by motion impairments, and it is relevant to obtain a quantified measure of the quality of movement of a patient to follow his/her recovery. The sensor signals were processed by applying two approaches, time-domain features and similarity distance between each pair of signals. The data analysis was divided into classification and variability using features and distances calculated previously. The classification analysis was made to determine if the movements performed by the test subjects of both groups are distinguishable among them. The variability analysis was conducted to measure the similarity of the movements. According to the results, the flexion/extension movement had a high intra-group variability. In addition, meaningful information were provided in terms of change of velocity and rotational motions for each individual.