Long-term activity recognition from wristwatch accelerometer data

Comparison of Students' Physical Activity at Different Times and Establishment of a Regression Model for Smart Fitness Trackers

Under the strategy of Healthy China, students' physical health status not only affects their future life and studies but also influences social progress and development. By monitoring and measuring the daily PA levels of Chinese students over a week, this study aimed to fully understand the current PA status of students at different times, providing data support for improving students' PA levels and physical health. (1) Wearable fitness trackers have advantages such as low cost, portable wearability, and intuitive test data. By exploring the differences between wearable devices and PA testing instruments, this study provides reference data to improve the accuracy of wearable devices and promote the use of fitness trackers instead of triaxial accelerometers, thereby advancing scientific research on PA and the development of mass fitness. A total of 261 students (147 males; 114 females) were randomly selected and wore both the Actigraph GT3X+ triaxial accelerometer and Huawei smart fitness trackers simultaneously to monitor their daily PA levels, energy metabolism, sedentary behavior, and step counts from the trackers over a week. The students' PA status and living habits were also understood through literature reviews and questionnaire surveys. The validity of the smart fitness trackers was quantitatively analyzed using ActiLife software 6 Data Analysis Software and traditional analysis methods such as MedCal. Paired sample Wilcoxon signed-rank tests and mean absolute error ratio tests were used to assess the validity of the smart fitness trackers relative to the Actigraph GT3X+ triaxial accelerometer. A linear regression model was established to predict the step counts of the Actigraph GT3X+ triaxial accelerometer based on the step counts from the smart fitness trackers, aiming to improve the accuracy of human motion measurement by smart fitness trackers. There were significant differences in moderate-to-high-intensity PA time, energy expenditure, metabolic equivalents, and step counts between males and females (p < 0.01), with females having higher values than males in both moderate-to-high-intensity PA time and step counts. Sedentary behavior showed significant differences only on weekdays between males and females (p < 0.05), with females engaging in less sedentary behavior than males. (2) There was a significant difference in sedentary time between weekdays and weekends for students (p < 0.05), with sedentary time being higher on weekends than on weekdays. (3) Compared with weekends, female students had significantly different moderate-to-high-intensity PA time and sedentary time on weekdays (p < 0.01), while no significant differences were observed for male students. (4) Under free-living conditions, the average daily step count monitored by the smart fitness trackers was lower than that measured by the Actigraph GT3X+ triaxial accelerometer, with a significant difference (p < 0.01), but both showed a positive correlation (r = 0.727). (5) The linear regression equation established between the step counts monitored by the smart fitness trackers and those by the Actigraph GT3X+ triaxial accelerometer was y = 3677.3157 + 0.6069x. The equation's R2 = 0.625, with an F-test value of p < 0.001, indicating a high degree of fit between the step counts recorded by the Huawei fitness tracker and those recorded by the triaxial accelerometer. The t-test results for the regression coefficient and constant term were t = 26.4410 and p < 0.01, suggesting that both were meaningful. The tested students were able to meet the recommended total amount of moderate-intensity PA for 150 min per week or high-intensity PA for 75 min per week according to the "Chinese Adult PA Guidelines", as well as the recommended daily step count of more than 6000 steps per day according to the "Chinese Dietary Guidelines". (2) Female students had significantly more moderate-to-high-intensity PA time than male students, but lower energy expenditure and metabolic equivalents, which may have been related to their lifestyle and types of exercise. On weekends, female students significantly increased their moderate-to-high-intensity PA time compared with males but also showed increased sedentary time exceeding that of males; further investigation is needed to understand the reasons behind these findings. (3) The step counts monitored by the Huawei smart fitness trackers correlated with those measured by the Actigraph GT3X+ triaxial accelerometer, but the step counts from the fitness trackers were lower, indicating that the fitness trackers underestimated PA levels. (4) There was a linear relationship between the Huawei smart fitness trackers and the Actigraph GT3X+ triaxial accelerometer. By using the step counts monitored by the Huawei fitness trackers and the regression equation, it was possible to estimate the activity counts from the Actigraph GT3X+ triaxial accelerometer. Replacing the Actigraph GT3X+ triaxial accelerometer with Huawei smart fitness trackers for step count monitoring significantly reduces testing costs while providing consumers with intuitive data.

An LSTM-based Gesture-to-Speech Recognition System

Fast and flexible communication options are limited for speech-impaired people. Hand gestures coupled with fast, generated speech can enable a more natural social dynamic for those individuals - particularly individuals without the fine motor skills to type on a keyboard or tablet reliably. We created a mobile phone application prototype that generates audible responses associated with trained hand movements and collects and organizes the accelerometer data for rapid training to allow tailored models for individuals who may not be able to perform standard movements such as sign language. Six participants performed 11 distinct gestures to produce the dataset. A mobile application was developed that integrated a bidirectional LSTM network architecture which was trained from this data. After evaluation using nested subject-wise cross-validation, our integrated bidirectional LSTM model demonstrates an overall recall of 91.8% in recognition of these pre-selected 11 hand gestures, with recall at 95.8% when two commonly confused gestures were not assessed. This prototype is a step in creating a mobile phone system capable of capturing new gestures and developing tailored gesture recognition models for individuals in speech-impaired populations. Further refinement of this prototype can enable fast and efficient communication with the goal of further improving social interaction for individuals unable to speak.

Physiotherapy Exercise Classification with Single-Camera Pose Detection and Machine Learning

Access to healthcare, including physiotherapy, is increasingly occurring through virtual formats. At-home adherence to physical therapy programs is often poor and few tools exist to objectively measure participation. The aim of this study was to develop and evaluate the potential for performing automatic, unsupervised video-based monitoring of at-home low-back and shoulder physiotherapy exercises using a mobile phone camera. Joint locations were extracted from the videos of healthy subjects performing low-back and shoulder physiotherapy exercises using an open source pose detection framework. A convolutional neural network was trained to classify physiotherapy exercises based on the segments of keypoint time series data. The model's performance as a function of input keypoint combinations was studied in addition to its robustness to variation in the camera angle. The CNN model achieved optimal performance using a total of 12 pose estimation landmarks from the upper and lower body (low-back exercise classification: 0.995 ± 0.009; shoulder exercise classification: 0.963 ± 0.020). Training the CNN on a variety of angles was found to be effective in making the model robust to variations in video filming angle. This study demonstrates the feasibility of using a smartphone camera and a supervised machine learning model to effectively classify at-home physiotherapy participation and could provide a low-cost, scalable method for tracking adherence to physical therapy exercise programs in a variety of settings.

A Multimodal Data Processing System for LiDAR-Based Human Activity Recognition

Increasingly, the task of detecting and recognizing the actions of a human has been delegated to some form of neural network processing camera or wearable sensor data. Due to the degree to which the camera can be affected by lighting and wearable sensors scantiness, neither one modality can capture the required data to perform the task confidently. That being the case, range sensors, like light detection and ranging (LiDAR), can complement the process to perceive the environment more robustly. Most recently, researchers have been exploring ways to apply convolutional neural networks to 3-D data. These methods typically rely on a single modality and cannot draw on information from complementing sensor streams to improve accuracy. This article proposes a framework to tackle human activity recognition by leveraging the benefits of sensor fusion and multimodal machine learning. Given both RGB and point cloud data, our method describes the activities being performed by subjects using regions with a convolutional neural network (R-CNN) and a 3-D modified Fisher vector network. Evaluated on a custom captured multimodal dataset demonstrates that the model outputs remarkably accurate human activity classification (90%). Furthermore, this framework can be used for sports analytics, understanding social behavior, surveillance, and perhaps most notably by autonomous vehicles (AVs) to data-driven decision-making policies in urban areas and indoor environments.

A machine learning approach for semi-automatic assessment of IADL dependence in older adults with wearable sensors

BACKGROUND AND OBJECTIVE

The assessment of dependence in older adults currently requires a manual collection of data taken from questionnaires. This process is time consuming for the clinicians and intrudes the daily life of the elderly. This paper aims to semi-automate the acquisition and analysis of health data to assess and predict the dependence in older adults while executing one instrumental activity of daily living (IADL).

METHODS

In a mobile-health (m-health) scenario, we analyze whether the acquisition of data through wearables during the performance of IADLs, and with the help of machine learning techniques could replace the traditional questionnaires to evaluate dependence. To that end, we collected data from wearables, while older adults do the shopping activity. A trial supervisor (TS) labelled the different shopping stages (SS) in the collected data. We performed data pre-processing techniques over those SS and analyzed them with three machine learning algorithms: k-Nearest Neighbors (k-NN), Random Forest (RF) and Support Vector Machines (SVM).

RESULTS

Our results confirm that it is possible to replace the traditional questionnaires with wearable data. In particular, the best learning algorithm we tried reported an accuracy of 97% in the assessment of dependence. We tuned the hyperparameters of this algorithm and used embedded feature selection technique to get the best performance with a subset of only 10 features out of the initial 85. This model considers only features extracted from four sensors of a single wearable: accelerometer, heart rate, electrodermal activity and temperature. Although these features are not observational, our current proposal is semi-automatic, because it needs a TS labelling the SS (with a smartphone application). In the future, this labelling process could be automatic as well.

CONCLUSIONS

Our method can semi-automatically assess the dependence, without disturbing daily activities of elderly people. This method can save clinicians' time in the evaluation of dependence in older adults and reduce healthcare costs.

Recognition of Bathroom Activities in Older Adults Using Wearable Sensors: A Systematic Review and Recommendations

This article provides a systematic review of studies on recognising bathroom activities in older adults using wearable sensors. Bathroom activities are an important part of Activities of Daily Living (ADL). The performance on ADL activities is used to predict the ability of older adults to live independently. This paper aims to provide an overview of the studied bathroom activities, the wearable sensors used, different applied methodologies and the tested activity recognition techniques. Six databases were screened up to March 2020, based on four categories of keywords: older adults, activity recognition, bathroom activities and wearable sensors. In total, 4262 unique papers were found, of which only seven met the inclusion criteria. This small number shows that few studies have been conducted in this field. Therefore, in addition, this critical review resulted in several recommendations for future studies. In particular, we recommend to (1) study complex bathroom activities, including multiple movements; (2) recruit participants, especially the target population; (3) conduct both lab and real-life experiments; (4) investigate the optimal number and positions of wearable sensors; (5) choose a suitable annotation method; (6) investigate deep learning models; (7) evaluate the generality of classifiers; and (8) investigate both detection and quality performance of an activity.

Out-of-Distribution Detection of Human Activity Recognition with Smartwatch Inertial Sensors

Out-of-distribution (OOD) in the context of Human Activity Recognition (HAR) refers to data from activity classes that are not represented in the training data of a Machine Learning (ML) algorithm. OOD data are a challenge to classify accurately for most ML algorithms, especially deep learning models that are prone to overconfident predictions based on in-distribution (IIN) classes. To simulate the OOD problem in physiotherapy, our team collected a new dataset (SPARS9x) consisting of inertial data captured by smartwatches worn by 20 healthy subjects as they performed supervised physiotherapy exercises (IIN), followed by a minimum 3 h of data captured for each subject as they engaged in unrelated and unstructured activities (OOD). In this paper, we experiment with three traditional algorithms for OOD-detection using engineered statistical features, deep learning-generated features, and several popular deep learning approaches on SPARS9x and two other publicly-available human activity datasets (MHEALTH and SPARS). We demonstrate that, while deep learning algorithms perform better than simple traditional algorithms such as KNN with engineered features for in-distribution classification, traditional algorithms outperform deep learning approaches for OOD detection for these HAR time series datasets.

An Integrated ARMA-Based Deep Autoencoder and GRU Classifier System for Enhanced Recognition of Daily Hand Activities

Recognition of hand activities of daily living (hand-ADL) is useful in the areas of human–computer interactions, lifelogging, and healthcare applications. However, developing a reliable human activity recognition (HAR) system for hand-ADL with only a single wearable sensor is still a challenge due to hand movements that are typically transient and sporadic. Approaches based on deep learning methodologies to reduce noise and extract relevant features directly from raw data are becoming more promising for implementing such HAR systems. In this work, we present an ARMA-based deep autoencoder and a deep recurrent network (RNN) using Gated Recurrent Unit (GRU) for recognition of hand-ADL using signals from a single IMU wearable sensor. The integrated ARMA-based autoencoder denoises raw time-series signals of hand activities, such that better representation of human hand activities can be made. Then, our deep RNN-GRU recognizes seven hand-ADL based upon the output of the autoencoder: namely, Open Door, Close Door, Open Refrigerator, Close Refrigerator, Open Drawer, Close Drawer, and Drink from Cup. The proposed methodology using RNN-GRU with autoencoder achieves a mean accuracy of 84.94% and F1-score of 83.05% outperforming conventional classifiers such as RNN-LSTM, BRNN-LSTM, CNN, and Hybrid-RNNs by 4–10% higher in both accuracy and F1-score. The experimental results also showed the use of the autoencoder improves both the accuracy and F1-score of each conventional classifier by 12.8% in RNN-LSTM, 4.37% in BRNN-LSTM, 15.45% CNN, 14.6% Hybrid RNN, and 12.4% for the proposed RNN-GRU.

Applying machine learning in motor activity time series of depressed bipolar and unipolar patients compared to healthy controls

Current practice of assessing mood episodes in affective disorders largely depends on subjective observations combined with semi-structured clinical rating scales. Motor activity is an objective observation of the inner physiological state expressed in behavior patterns. Alterations of motor activity are essential features of bipolar and unipolar depression. The aim was to investigate if objective measures of motor activity can aid existing diagnostic practice, by applying machine-learning techniques to analyze activity patterns in depressed patients and healthy controls. Random Forrest, Deep Neural Network and Convolutional Neural Network algorithms were used to analyze 14 days of actigraph recorded motor activity from 23 depressed patients and 32 healthy controls. Statistical features analyzed in the dataset were mean activity, standard deviation of mean activity and proportion of zero activity. Various techniques to handle data imbalance were applied, and to ensure generalizability and avoid overfitting a Leave-One-User-Out validation strategy was utilized. All outcomes reports as measures of accuracy for binary tests. A Deep Neural Network combined with SMOTE class balancing technique performed a cut above the rest with a true positive rate of 0.82 (sensitivity) and a true negative rate of 0.84 (specificity). Accuracy was 0.84 and the Matthews Correlation Coefficient 0.65. Misclassifications appear related to data overlapping among the classes, so an appropriate future approach will be to compare mood states intra-individualistically. In summary, machine-learning techniques present promising abilities in discriminating between depressed patients and healthy controls in motor activity time series.

A Microservices e-Health System for Ecological Frailty Assessment Using Wearables

The population in developed countries is aging and this fact results in high elderly health costs, as well as a decrease in the number of active working members to support these costs. This could lead to a collapse of the current systems. One of the first insights of the decline in elderly people is frailty, which could be decelerated if it is detected at an early stage. Nowadays, health professionals measure frailty manually through questionnaires and tests of strength or gait focused on the physical dimension. Sensors are increasingly used to measure and monitor different e-health indicators while the user is performing Basic Activities of Daily Life (BADL). In this paper, we present a system based on microservices architecture, which collects sensory data while the older adults perform Instrumental ADLs (IADLs) in combination with BADLs. IADLs involve physical dimension, but also cognitive and social dimensions. With the sensory data we built a machine learning model to assess frailty status which outperforms the previous works that only used BADLs. Our model is accurate, ecological, non-intrusive, flexible and can help health professionals to automatically detect frailty.

Multi-Sensor Fusion for Activity Recognition-A Survey

In Ambient Intelligence (AmI), the activity a user is engaged in is an essential part of the context, so its recognition is of paramount importance for applications in areas like sports, medicine, personal safety, and so forth. The concurrent use of multiple sensors for recognition of human activities in AmI is a good practice because the information missed by one sensor can sometimes be provided by the others and many works have shown an accuracy improvement compared to single sensors. However, there are many different ways of integrating the information of each sensor and almost every author reporting sensor fusion for activity recognition uses a different variant or combination of fusion methods, so the need for clear guidelines and generalizations in sensor data integration seems evident. In this survey we review, following a classification, the many fusion methods for information acquired from sensors that have been proposed in the literature for activity recognition; we examine their relative merits, either as they are reported and sometimes even replicated and a comparison of these methods is made, as well as an assessment of the trends in the area.

Human Activity Recognition for Production and Logistics—A Systematic Literature Review

This contribution provides a systematic literature review of Human Activity Recognition for Production and Logistics. An initial list of 1243 publications that complies with predefined Inclusion Criteria was surveyed by three reviewers. Fifty-two publications that comply with the Content Criteria were analysed regarding the observed activities, sensor attachment, utilised datasets, sensor technology and the applied methods of HAR. This review is focused on applications that use marker-based Motion Capturing or Inertial Measurement Units. The analysed methods can be deployed in industrial application of Production and Logistics or transferred from related domains into this field. The findings provide an overview of the specifications of state-of-the-art HAR approaches, statistical pattern recognition and deep architectures and they outline a future road map for further research from a practitioner’s perspective.

Wearable Sensor-Based Human Activity Recognition via Two-Layer Diversity-Enhanced Multiclassifier Recognition Method

Sensor-based human activity recognition can benefit a variety of applications such as health care, fitness, smart homes, rehabilitation training, and so forth. In this paper, we propose a novel two-layer diversity-enhanced multiclassifier recognition method for single wearable accelerometer-based human activity recognition, which contains data-based and classifier-based diversity enhancement. Firstly, we introduce the kernel Fisher discriminant analysis (KFDA) technique to spatially transform the training samples and enhance the discrimination between activities. In addition, bootstrap resampling is utilized to increase the diversities of the dataset for training the base classifiers in the multiclassifier system. Secondly, a combined diversity measure for selecting the base classifiers with excellent performance and large diversity is proposed to optimize the performance of the multiclassifier system. Lastly, majority voting is utilized to combine the preferred base classifiers. Experiments showed that the data-based diversity enhancement can improve the discriminance of different activity samples and promote the generation of base classifiers with different structures and performances. Compared with random selection and traditional ensemble methods, including Bagging and Adaboost, the proposed method achieved 92.3% accuracy and 90.7% recall, which demonstrates better performance in activity recognition.

Deep learning-based classification with improved time resolution for physical activities of children

Background

The proportion of overweight and obese people has increased tremendously in a short period, culminating in a worldwide trend of obesity that is reaching epidemic proportions. Overweight and obesity are serious issues, especially with regard to children. This is because obese children have twice the risk of becoming obese as adults, as compared to non-obese children. Nowadays, many methods for maintaining a caloric balance exist; however, these methods are not applicable to children. In this study, a new approach for helping children monitor their activities using a convolutional neural network (CNN) is proposed, which is applicable for real-time scenarios requiring high accuracy.

Methods

A total of 136 participants (86 boys and 50 girls), aged between 8.5 years and 12.5 years (mean 10.5, standard deviation 1.1), took part in this study. The participants performed various movement while wearing custom-made three-axis accelerometer modules around their waists. The data acquired by the accelerometer module was preprocessed by dividing them into small sets (128 sample points for 2.8 s). Approximately 183,600 data samples were used by the developed CNN for learning to classify ten physical activities : slow walking, fast walking, slow running, fast running, walking up the stairs, walking down the stairs, jumping rope, standing up, sitting down, and remaining still.

Results

The developed CNN classified the ten activities with an overall accuracy of 81.2%. When similar activities were merged, leading to seven merged activities, the CNN classified activities with an overall accuracy of 91.1%. Activity merging also improved performance indicators, for the maximum case of 66.4% in recall, 48.5% in precision, and 57.4% in f1 score . The developed CNN classifier was compared to conventional machine learning algorithms such as the support vector machine, decision tree, and k-nearest neighbor algorithms, and the proposed CNN classifier performed the best: CNN (81.2%) > SVM (64.8%) > DT (63.9%) > kNN (55.4%) (for ten activities); CNN (91.1%) > SVM (74.4%) > DT (73.2%) > kNN (65.3%) (for the merged seven activities).

Discussion

The developed algorithm distinguished physical activities with improved time resolution using short-time acceleration signals from the physical activities performed by children. This study involved algorithm development, participant recruitment, IRB approval, custom-design of a data acquisition module, and data collection. The self-selected moving speeds for walking and running (slow and fast) and the structure of staircase degraded the performance of the algorithm. However, after similar activities were merged, the effects caused by the self-selection of speed were reduced. The experimental results show that the proposed algorithm performed better than conventional algorithms. Owing to its simplicity, the proposed algorithm could be applied to real-time applicaitons.

Shoulder physiotherapy exercise recognition: machine learning the inertial signals from a smartwatch

OBJECTIVE

Participation in a physical therapy program is considered one of the greatest predictors of successful conservative management of common shoulder disorders. However, adherence to these protocols is often poor and typically worse for unsupervised home exercise programs. Currently, there are limited tools available for objective measurement of adherence in the home setting. The goal of this study was to develop and evaluate the potential for performing home shoulder physiotherapy monitoring using a commercial smartwatch.

APPROACH

Twenty healthy adult subjects with no prior shoulder disorders performed seven exercises from an evidence-based rotator cuff physiotherapy protocol, while 6-axis inertial sensor data was collected from the active extremity. Within an activity recognition chain (ARC) framework, four supervised learning algorithms were trained and optimized to classify the exercises: k-nearest neighbor (k-NN), random forest (RF), support vector machine classifier (SVC), and a convolutional recurrent neural network (CRNN). Algorithm performance was evaluated using 5-fold cross-validation stratified first temporally and then by subject.

MAIN RESULTS

Categorical classification accuracy was above 94% for all algorithms on the temporally stratified cross validation, with the best performance achieved by the CRNN algorithm (99.4%). The subject stratified cross validation, which evaluated classifier performance on unseen subjects, yielded lower accuracies scores again with CRNN performing best (88.9%).

SIGNIFICANCE

This proof of concept study demonstrates the technical feasibility of a smartwatch device and supervised machine learning approach to more easily monitor and assess the at-home adherence of shoulder physiotherapy exercise protocols.

An Easily Customized Gesture Recognizer for Assisted Living Using Commodity Mobile Devices

Automatic gesture recognition is an important field in the area of human-computer interaction. Until recently, the main approach to gesture recognition was based mainly on real time video processing. The objective of this work is to propose the utilization of commodity smartwatches for such purpose. Smartwatches embed accelerometer sensors, and they are endowed with wireless communication capabilities (primarily Bluetooth), so as to connect with mobile phones on which gesture recognition algorithms may be executed. The algorithmic approach proposed in this paper accepts as the input readings from the smartwatch accelerometer sensors and processes them on the mobile phone. As a case study, the gesture recognition application was developed for Android devices and the Pebble smartwatch. This application allows the user to define the set of gestures and to train the system to recognize them. Three alternative methodologies were implemented and evaluated using a set of six 3-D natural gestures. All the reported results are quite satisfactory, while the method based on SAX (Symbolic Aggregate approXimation) was proven the most efficient.

An Activity Recognition Framework Deploying the Random Forest Classifier and A Single Optical Heart Rate Monitoring and Triaxial AccelerometerWrist-Band

Wrist-worn sensors have better compliance for activity monitoring compared to hip, waist, ankle or chest positions. However, wrist-worn activity monitoring is challenging due to the wide degree of freedom for the hand movements, as well as similarity of hand movements in different activities such as varying intensities of cycling. To strengthen the ability of wrist-worn sensors in detecting human activities more accurately, motion signals can be complemented by physiological signals such as optical heart rate (HR) based on photoplethysmography. In this paper, an activity monitoring framework using an optical HR sensor and a triaxial wrist-worn accelerometer is presented. We investigated a range of daily life activities including sitting, standing, household activities and stationary cycling with two intensities. A random forest (RF) classifier was exploited to detect these activities based on the wrist motions and optical HR. The highest overall accuracy of 89.6 ± 3.9% was achieved with a forest of a size of 64 trees and 13-s signal segments with 90% overlap. Removing the HR-derived features decreased the classification accuracy of high-intensity cycling by almost 7%, but did not affect the classification accuracies of other activities. A feature reduction utilizing the feature importance scores of RF was also carried out and resulted in a shrunken feature set of only 21 features. The overall accuracy of the classification utilizing the shrunken feature set was 89.4 ± 4.2%, which is almost equivalent to the above-mentioned peak overall accuracy.

[-25]A Similarity Analysis of Audio Signal to Develop a Human Activity Recognition Using Similarity Networks

Human Activity Recognition (HAR) is one of the main subjects of study in the areas of computer vision and machine learning due to the great benefits that can be achieved. Examples of the study areas are: health prevention, security and surveillance, automotive research, and many others. The proposed approaches are carried out using machine learning techniques and present good results. However, it is difficult to observe how the descriptors of human activities are grouped. In order to obtain a better understanding of the the behavior of descriptors, it is important to improve the abilities to recognize the human activities. This paper proposes a novel approach for the HAR based on acoustic data and similarity networks. In this approach, we were able to characterize the sound of the activities and identify those activities looking for similarity in the sound pattern. We evaluated the similarity of the sounds considering mainly two features: the sound location and the materials that were used. As a result, the materials are a good reference classifying the human activities compared with the location.

An Improved Three-Stage Classifier for Activity Recognition

Recently, Human Activity Recognition (HAR) has become an important research area because of its wide range of applications in several domains such as health care, elder care, sports monitoring systems, etc. The use of wearable sensors — specifically the use of inertial sensors such as accelerometers and gyroscopes — has become the most common approach to recognize physical activities because of their unobtrusiveness and ubiquity. Overall, the process of building a HAR system starts with a feature extraction phase and then a classification model is trained. In the work of Siirtola et al. is proposed an intermediate clustering step to find the homogeneous groups of activities. For the recognition step, an instance is assigned to one of the groups and the final classification is performed inside that group. In this work we evaluate the clustering-based approach for activity classification proposed by Siirtola with two additional improvements: automatic selection of the number of groups and an instance reassignment procedure. In the original work, they evaluated their method using decision trees on a sports activities dataset. For our experiments, we evaluated seven different classification models on four public activity recognition datasets. Our results with 10-fold Cross Validation showed that the method proposed by Siirtola with our additional two improvements performed better in the majority of cases as compared to using the single classification model under consideration. When using Leave One User Out Cross Validation (user independent model) we found no differences between the proposed method and the single classification model.

A Behaviour Monitoring System (BMS) for Ambient Assisted Living

Unusual changes in the regular daily mobility routine of an elderly person at home can be an indicator or early symptom of developing health problems. Sensor technology can be utilised to complement the traditional healthcare systems to gain a more detailed view of the daily mobility of a person at home when performing everyday tasks. We hypothesise that data collected from low-cost sensors such as presence and occupancy sensors can be analysed to provide insights on the daily mobility habits of the elderly living alone at home and to detect routine changes. We validate this hypothesis by designing a system that automatically learns the daily room-to-room transitions and permanence habits in each room at each time of the day and generates alarm notifications when deviations are detected. We present an algorithm to process the sensors' data streams and compute sensor-driven features that describe the daily mobility routine of the elderly as part of the developed Behaviour Monitoring System (BMS). We are able to achieve low detection delay with confirmation time that is high enough to convey the detection of a set of common abnormal situations. We illustrate and evaluate BMS with synthetic data, generated by a developed data generator that was designed to mimic different user's mobility profiles at home, and also with a real-life dataset collected from prior research work. Results indicate BMS detects several mobility changes that can be symptoms of common health problems. The proposed system is a useful approach for learning the mobility habits at the home environment, with the potential to detect behaviour changes that occur due to health problems, and therefore, motivating progress toward behaviour monitoring and elder's care.

Daily wrist activity classification using a smart band

OBJECTIVE

In this letter, we propose a novel method for classifying daily wrist activities by using a smart band.

APPROACH

Triaxial acceleration data are collected by built-in sensors of the smart band during experiments regarding five activities, i.e. texting, calling, placing a hand in a pocket, carrying a suitcase, and swinging a hand. We analyze patterns in the sensor signals during these activities based on three types of features, i.e. norm, norm-variance, and frequency-domain features. After extracting the significant features, a multi-class support vector machine algorithm is applied to classify these activities.

MAIN RESULTS

We obtained recognition error rates of approximately 2.7% by applying the proposed method to the experimental dataset.

Location-Enhanced Activity Recognition in Indoor Environments Using Off the Shelf Smart Watch Technology and BLE Beacons

Activity recognition in indoor spaces benefits context awareness and improves the efficiency of applications related to personalised health monitoring, building energy management, security and safety. The majority of activity recognition frameworks, however, employ a network of specialised building sensors or a network of body-worn sensors. As this approach suffers with respect to practicality, we propose the use of commercial off-the-shelf devices. In this work, we design and evaluate an activity recognition system composed of a smart watch, which is enhanced with location information coming from Bluetooth Low Energy (BLE) beacons. We evaluate the performance of this approach for a variety of activities performed in an indoor laboratory environment, using four supervised machine learning algorithms. Our experimental results indicate that our location-enhanced activity recognition system is able to reach a classification accuracy ranging from 92% to 100%, while without location information classification accuracy it can drop to as low as 50% in some cases, depending on the window size chosen for data segmentation.

Hierarchical Activity Recognition Using Smart Watches and RGB-Depth Cameras

Human activity recognition is important for healthcare and lifestyle evaluation. In this paper, a novel method for activity recognition by jointly considering motion sensor data recorded by wearable smart watches and image data captured by RGB-Depth (RGB-D) cameras is presented. A normalized cross correlation based mapping method is implemented to establish association between motion sensor data with corresponding image data from the same person in multi-person situations. Further, to improve the performance and accuracy of recognition, a hierarchical structure embedded with an automatic group selection method is proposed. Through this method, if the number of activities to be classified is changed, the structure will be changed correspondingly without interaction. Our comparative experiments against the single data source and single layer methods have shown that our method is more accurate and robust.

Actigraphy features for predicting mobility disability in older adults

Actigraphy has attracted much attention for assessing physical activity in the past decade. Many algorithms have been developed to automate the analysis process, but none has targeted a general model to discover related features for detecting or predicting mobility function, or more specifically, mobility impairment and major mobility disability (MMD). Men (N  =  357) and women (N  =  778) aged 70-89 years wore a tri-axial accelerometer (Actigraph GT3X) on the right hip during free-living conditions for 8.4  ±  3.0 d. One-second epoch data were summarized into 67 features. Several machine learning techniques were used to select features from the free-living condition to predict mobility impairment, defined as 400 m walking speed  <0.80 m s^-1. Selected features were also included in a model to predict the first occurrence of MMD-defined as the loss in the ability to walk 400 m. Each method yielded a similar estimate of 400 m walking speed with a root mean square error of ~0.07 m s^-1 and an R-squared values ranging from 0.37-0.41. Sensitivity and specificity of identifying slow walkers was approximately 70% and 80% for all methods, respectively. The top five features, which were related to movement pace and amount (activity counts and steps), length in activity engagement (bout length), accumulation patterns of activity, and movement variability significantly improved the prediction of MMD beyond that found with common covariates (age, diseases, anthropometry, etc). This study identified a subset of actigraphy features collected in free-living conditions that are moderately accurate in identifying persons with clinically-assessed mobility impaired and significantly improve the prediction of MMD. These findings suggest that the combination of features as opposed to a specific feature is important to consider when choosing features and/or combinations of features for prediction of mobility phenotypes in older adults.

Progress in Biomedical Knowledge Discovery: A 25-year Retrospective

OBJECTIVES

We sought to explore, via a systematic review of the literature, the state of the art of knowledge discovery in biomedical databases as it existed in 1992, and then now, 25 years later, mainly focused on supervised learning.

METHODS

We performed a rigorous systematic search of PubMed and latent Dirichlet allocation to identify themes in the literature and trends in the science of knowledge discovery in and between time periods and compare these trends. We restricted the result set using a bracket of five years previous, such that the 1992 result set was restricted to articles published between 1987 and 1992, and the 2015 set between 2011 and 2015. This was to reflect the current literature available at the time to researchers and others at the target dates of 1992 and 2015. The search term was framed as: Knowledge Discovery OR Data Mining OR Pattern Discovery OR Pattern Recognition, Automated.

RESULTS

A total 538 and 18,172 documents were retrieved for 1992 and 2015, respectively. The number and type of data sources increased dramatically over the observation period, primarily due to the advent of electronic clinical systems. The period 1992- 2015 saw the emergence of new areas of research in knowledge discovery, and the refinement and application of machine learning approaches that were nascent or unknown in 1992.

CONCLUSIONS

Over the 25 years of the observation period, we identified numerous developments that impacted the science of knowledge discovery, including the availability of new forms of data, new machine learning algorithms, and new application domains. Through a bibliometric analysis we examine the striking changes in the availability of highly heterogeneous data resources, the evolution of new algorithmic approaches to knowledge discovery, and we consider from legal, social, and political perspectives possible explanations of the growth of the field. Finally, we reflect on the achievements of the past 25 years to consider what the next 25 years will bring with regard to the availability of even more complex data and to the methods that could be, and are being now developed for the discovery of new knowledge in biomedical data.

The Evaluation of Physical Stillness with Wearable Chest and Arm Accelerometer during Chan Ding Practice

Chan Ding training is beneficial to health and emotional wellbeing. More and more people have taken up this practice over the past few years. A major training method of Chan Ding is to focus on the ten Mailuns, i.e., energy points, and to maintain physical stillness. In this article, wireless wearable accelerometers were used to detect physical stillness, and the created physical stillness index (PSI) was also shown. Ninety college students participated in this study. Primarily, accelerometers used on the arms and chest were examined. The results showed that the PSI values on the arms were higher than that of the chest, when participants moved their bodies in three different ways, left-right, anterior-posterior, and hand, movements with natural breathing. Then, they were divided into three groups to practice Chan Ding for approximately thirty minutes. Participants without any Chan Ding experience were in Group I. Participants with one year of Chan Ding experience were in Group II, and participants with over three year of experience were in Group III. The Chinese Happiness Inventory (CHI) was also conducted. Results showed that the PSI of the three groups measured during 20-30 min were 0.123 ± 0.155, 0.012 ± 0.013, and 0.001 ± 0.0003, respectively (p < 0.001 ***). The averaged CHI scores of the three groups were 10.13, 17.17, and 25.53, respectively (p < 0.001 ***). Correlation coefficients between PSI and CHI of the three groups were -0.440, -0.369, and -0.537, respectively (p < 0.01 **). PSI value and the wearable accelerometer that are presently available on the market could be used to evaluate the quality of the physical stillness of the participants during Chan Ding practice.

Activity Recognition Using Community Data to Complement Small Amounts of Labeled Instances

Human Activity Recognition (HAR) is an important part of ambient intelligence systems since it can provide user-context information, thus allowing a greater personalization of services. One of the problems with HAR systems is that the labeling process for the training data is costly, which has hindered its practical application. A common approach is to train a general model with the aggregated data from all users. The problem is that for a new target user, this model can perform poorly because it is biased towards the majority type of users and does not take into account the particular characteristics of the target user. To overcome this limitation, a user-dependent model can be trained with data only from the target user that will be optimal for this particular user; however, this requires a considerable amount of labeled data, which is cumbersome to obtain. In this work, we propose a method to build a personalized model for a given target user that does not require large amounts of labeled data. Our method uses data already labeled by a community of users to complement the scarce labeled data of the target user. Our results showed that the personalized model outperformed the general and the user-dependent models when labeled data is scarce.

How Accurately Can Your Wrist Device Recognize Daily Activities and Detect Falls?

Although wearable accelerometers can successfully recognize activities and detect falls, their adoption in real life is low because users do not want to wear additional devices. A possible solution is an accelerometer inside a wrist device/smartwatch. However, wrist placement might perform poorly in terms of accuracy due to frequent random movements of the hand. In this paper we perform a thorough, large-scale evaluation of methods for activity recognition and fall detection on four datasets. On the first two we showed that the left wrist performs better compared to the dominant right one, and also better compared to the elbow and the chest, but worse compared to the ankle, knee and belt. On the third (Opportunity) dataset, our method outperformed the related work, indicating that our feature-preprocessing creates better input data. And finally, on a real-life unlabeled dataset the recognized activities captured the subject's daily rhythm and activities. Our fall-detection method detected all of the fast falls and minimized the false positives, achieving 85% accuracy on the first dataset. Because the other datasets did not contain fall events, only false positives were evaluated, resulting in 9 for the second, 1 for the third and 15 for the real-life dataset (57 days data).

Physical Human Activity Recognition Using Wearable Sensors

This paper presents a review of different classification techniques used to recognize human activities from wearable inertial sensor data. Three inertial sensor units were used in this study and were worn by healthy subjects at key points of upper/lower body limbs (chest, right thigh and left ankle). Three main steps describe the activity recognition process: sensors’ placement, data pre-processing and data classification. Four supervised classification techniques namely, k-Nearest Neighbor (k-NN), Support Vector Machines (SVM), Gaussian Mixture Models (GMM), and Random Forest (RF) as well as three unsupervised classification techniques namely, k-Means, Gaussian mixture models (GMM) and Hidden Markov Model (HMM), are compared in terms of correct classification rate, F-measure, recall, precision, and specificity. Raw data and extracted features are used separately as inputs of each classifier. The feature selection is performed using a wrapper approach based on the RF algorithm. Based on our experiments, the results obtained show that the k-NN classifier provides the best performance compared to other supervised classification algorithms, whereas the HMM classifier is the one that gives the best results among unsupervised classification algorithms. This comparison highlights which approach gives better performance in both supervised and unsupervised contexts. It should be noted that the obtained results are limited to the context of this study, which concerns the classification of the main daily living human activities using three wearable accelerometers placed at the chest, right shank and left ankle of the subject.