Search

The electroencephalogram (EEG) signal is a key parameter used to identify the different sleep stages present in an overnight sleep recording. Sleep staging is crucial in the diagnosis of several sleep disorders; however, the manual annotation of the EEG signal is a costly and time-consuming process. Automatic sleep staging algorithms offer a practical and cost-effective alternative to manual sleep staging. However, due to the limited availability of EEG sleep datasets, the reliability of existing sleep staging algorithms is questionable. Furthermore, most reported experimental results have been obtained using adult EEG signals; the effectiveness of these algorithms using pediatric EEGs is unknown. In this paper, we conduct an intensive study of two state-of-the-art single-channel EEG-based sleep staging algorithms, namely DeepSleepNet and AttnSleep, using a recently released large-scale sleep dataset collected from 3984 patients, most of whom are children. The paper studies how the performance of these sleep staging algorithms varies when applied on different EEG channels and across different age groups. Furthermore, all results were analyzed within individual sleep stages to understand how each stage is affected by the choice of EEG channel and the participants' age. The study concluded that the selection of the channel is crucial for the accuracy of the single-channel EEG-based automatic sleep staging methods. For instance, channels O1-M2 and O2-M1 performed consistently worse than other channels for both algorithms and through all age groups. The study also revealed the challenges in the automatic sleep staging of newborns and infants (1–52 weeks). [ABSTRACT FROM AUTHOR]

In this paper, we face the problem of task classification starting from physiological signals acquired using wearable sensors with experiments in a controlled environment, designed to consider two different age populations: young adults and older adults. Two different scenarios are considered. In the first one, subjects are involved in different cognitive load tasks, while in the second one, space varying conditions are considered, and subjects interact with the environment, changing the walking conditions and avoiding collision with obstacles. Here, we demonstrate that it is possible not only to define classifiers that rely on physiological signals to predict tasks that imply different cognitive loads, but it is also possible to classify both the population group age and the performed task. The whole workflow of data collection and analysis, starting from the experimental protocol, data acquisition, signal denoising, normalization with respect to subject variability, feature extraction and classification is described here. The dataset collected with the experiments together with the codes to extract the features of the physiological signals are made available for the research community. [ABSTRACT FROM AUTHOR]

Large-scale functional connectivity is an important indicator of the brain's normal functioning. The abnormalities in the connectivity pattern can be used as a diagnostic tool to detect various neurological disorders. The present paper describes the functional connectivity assessment based on artificial intelligence to reveal age-related changes in neural response in a simple motor execution task. Twenty subjects of two age groups performed repetitive motor tasks on command, while the whole-scalp EEG was recorded. We applied the model based on the feed-forward multilayer perceptron to detect functional relationships between five groups of sensors located over the frontal, parietal, left, right, and middle motor cortex. Functional dependence was evaluated with the predicted and original time series coefficient of determination. Then, we applied statistical analysis to highlight the significant features of the functional connectivity network assessed by our model. Our findings revealed the connectivity pattern is consistent with modern ideas of the healthy aging effect on neural activation. Elderly adults demonstrate a pronounced activation of the whole-brain theta-band network and decreased activation of frontal–parietal and motor areas of the mu-band. Between-subject analysis revealed a strengthening of inter-areal task-relevant links in elderly adults. These findings can be interpreted as an increased cognitive demand in elderly adults to perform simple motor tasks with the dominant hand, induced by age-related working memory decline. [ABSTRACT FROM AUTHOR]

Falls are unusual actions that cause a significant health risk among older people. The growing percentage of people of old age requires urgent development of fall detection and prevention systems. The emerging technology focuses on developing such systems to improve quality of life, especially for the elderly. A fall prevention system tries to predict and reduce the risk of falls. In contrast, a fall detection system observes the fall and generates a help notification to minimize the consequences of falls. A plethora of technical and review papers exist in the literature with a primary focus on fall detection. Similarly, several studies are relatively old, with a focus on wearables only, and use statistical and threshold-based approaches with a high false alarm rate. Therefore, this paper presents the latest research trends in fall detection and prevention systems using Machine Learning (ML) algorithms. It uses recent studies and analyzes datasets, age groups, ML algorithms, sensors, and location. Additionally, it provides a detailed discussion of the current trends of fall detection and prevention systems with possible future directions. This overview can help researchers understand the current systems and propose new methodologies by improving the highlighted issues. [ABSTRACT FROM AUTHOR]

Great toe strength (GTS) weakness is linked to declines in balance and mobility. Accurately assessing GTS, particularly great toe extension strength (GTES), is often neglected in clinical evaluations due to cumbersome and subjective methods. This study aims to characterize the force development curve output from the ToeScale and examine GTES variations with age, sex, BMI, and grip strength (GS) using traditional analyses and machine learning (ML). We conducted a pilot, cross-sectional feasibility study with convenience samples. We assessed GS using a hand-grip dynamometer and GTES using the ToeScale. The data analysis included descriptive statistics, correlations, independent samples t-tests, and accuracy and area under the curve (AUC) scores for three ML models. Thirty-one participants (males: 9; females: 22), 14 young (18–24 years) and 17 older (>65 years) adults, participated in the study. Males had significantly higher peak GTES than females in both age groups. The associations of GTES parameters with BMI and GS varied by age and sex. The ML model accuracies and AUC scores were low–moderate but aligned with traditional analyses. Future studies with larger samples and optimized ML models are needed. [ABSTRACT FROM AUTHOR]

For analytical approach-based word recognition techniques, the task of segmenting the word into individual characters is a big challenge, specifically for cursive handwriting. For this, a holistic approach can be a better option, wherein the entire word is passed to an appropriate recognizer. Gurumukhi script is a complex script for which a holistic approach can be proposed for offline handwritten word recognition. In this paper, the authors propose a Convolutional Neural Network-based architecture for recognition of the Gurumukhi month names. The architecture is designed with five convolutional layers and three pooling layers. The authors also prepared a dataset of 24,000 images, each with a size of 50 × 50. The dataset was collected from 500 distinct writers of different age groups and professions. The proposed method achieved training and validation accuracies of about 97.03% and 99.50%, respectively for the proposed dataset. [ABSTRACT FROM AUTHOR]

Heart rate variability (HRV) is related to cardiac vagal control and emotional regulation and an index for cardiac vagal control and cardiac autonomic activity. This study aimed to develop the Taiwan HRV normative database covering individuals aged 20 to 70 years and to assess its diagnosing validity in patients with major depressive disorder (MDD). A total of 311 healthy participants were in the HRV normative database and divided into five groups in 10-year age groups, and then the means and standard deviations of the HRV indices were calculated. We recruited 272 patients with MDD for cross-validation, compared their HRV indices with the normative database, and then converted them to Z-scores to explore the deviation of HRV in MDD patients from healthy groups. The results found a gradual decline in HRV indices with advancing age in the HC group, and females in the HC group exhibit higher cardiac vagal control and parasympathetic activity than males. Conversely, patients in the MDD group demonstrate lower HRV indices than those in the HC group, with their symptoms of depression and anxiety showing a negative correlation with HRV indices. The Taiwan HRV normative database has good psychometric characteristics of cross-validation. [ABSTRACT FROM AUTHOR]

The ongoing intense development of short-range radar systems and their improved capability of measuring small movements make these systems reliable solutions for the extraction of human vital signs in a contactless fashion. The continuous contactless monitoring of vital signs can be considered in a wide range of applications, such as remote healthcare solutions and context-aware smart sensor development. Currently, the provision of radar-recorded datasets of human vital signs is still an open issue. In this paper, we present a new frequency-modulated continuous wave (FMCW) radar-recorded vital sign dataset for 50 children aged less than 13 years. A clinically approved vital sign monitoring sensor was also deployed as a reference, and data from both sensors were time-synchronized. With the presented dataset, a new child age-group classification system based on GoogLeNet is proposed to develop a child safety sensor for smart vehicles. The radar-recorded vital signs of children are divided into several age groups, and the GoogLeNet framework is trained to predict the age of unknown human test subjects. [ABSTRACT FROM AUTHOR]

COVID-19 has disrupted every field of life and education is not immune to it. Student learning and examinations moved on-line on a few weeks notice, which has created a large workload for academics to grade the assessments and manually detect students' dishonesty. In this paper, we propose a method to automatically indicate cheating in unproctored on-line exams, when somebody else other than the legitimate student takes the exam. The method is based on the analysis of the student's on-line traces, which are logged by distance education systems. We work with customized IP geolocation and other data to derive the student's cheating risk score. We apply the method to approx. 3600 students in 22 courses, where the partial or final on-line exams were unproctored. The found cheating risk scores are presented along with examples of indicated cheatings. The method can be used to select students for knowledge re-validation, or to compare student cheating across courses, age groups, countries, and universities. We compared student cheating risk scores between four academic terms, including two terms of university closure due to COVID-19. [ABSTRACT FROM AUTHOR]

Due to the continuous monitoring process of critical patients, Intensive Care Units (ICU) generate large amounts of data, which are difficult for healthcare personnel to analyze manually, especially in overloaded situations such as those present during the COVID-19 pandemic. Therefore, the automatic analysis of these data has many practical applications in patient monitoring, including the optimization of alarm systems for alerting healthcare personnel. In this paper, explainable machine learning techniques are used for this purpose, with a methodology based on age-stratification, boosting classifiers, and Shapley Additive Explanations (SHAP) proposed. The methodology is evaluated using MIMIC-III, an ICU patient research database. The results show that the proposed model can predict mortality within the ICU with AUROC values of 0.961, 0.936, 0.898, and 0.883 for age groups 18–45, 45–65, 65–85 and 85+, respectively. By using SHAP, the features with the highest impact in predicting mortality for different age groups and the threshold from which the value of a clinical feature has a negative impact on the patient's health can be identified. This allows ICU alarms to be improved by identifying the most important variables to be sensed and the threshold values at which the health personnel must be warned. [ABSTRACT FROM AUTHOR]

The aim of this study was to determine if specific physical tests are sufficiently discriminant to differentiate players of similar anthropometric characteristics, but of different playing levels. Physical tests were conducted analyzing specific strength, throwing velocity, and running speed tests. Thirty-six male junior handball players (n = 36; age 19.7 ± 1.8 years; 185.6 ± 6.9 cm; 83.1 ± 10.3 kg; 10.6 ± 3.2 years of experience) from two different levels of competition participated in the study: NT = 18 were world top-level elite players, belonging to the Spanish junior men's national team (National Team = NT) and A = 18 players of the same age and anthropometric conditions, who were selected from Spanish third league men's teams (Amateur = A). The results showed significant differences (p < 0.05) between the two groups in all physical tests, except for two-step-test velocity and shoulder internal rotation. We conclude that a battery combining the Specific Performance Test and the Force Development Standing Test is useful in identifying talent and differentiating between elite and sub-elite players. The current findings suggest that running speed tests and throwing tests are essential in selecting players, regardless of age, sex, or type of competition. The results shed light on the factors that differentiate players of different levels and can help coaches in selecting players. [ABSTRACT FROM AUTHOR]

Joint quasi-stiffness has been often used to inform exoskeleton design. Further understanding of hip quasi-stiffness is needed to design hip exoskeletons. Of interest are wearer responses to walking speed changes with added mass of the exoskeleton. This study analyzed hip quasi-stiffness at 3 walking speed levels and 9 added mass distributions among 13 young and 16 middle-aged adults during mid-stance hip extension and late-stance hip flexion. Compared to young adults, middle-aged adults maintained a higher quasi-stiffness with a smaller range. For a faster walking speed, both age groups increased extension and flexion quasi-stiffness. With mass evenly distributed on the pelvis and thighs or biased to the pelvis, both groups maintained or increased extension quasi-stiffness. With mass biased to the thighs, middle-aged adults maintained or decreased extension quasi-stiffness while young adults increased it. Young adults decreased flexion quasi-stiffness with added mass but not in any generalizable pattern with mass amounts or distributions. Conversely, middle-aged adults maintained or decreased flexion quasi-stiffness with even distribution on the pelvis and thighs or biased to the pelvis, while no change occurred if biased to the thighs. In conclusion, these results can guide the design of a hip exoskeleton's size and mass distribution according to the intended user's age. [ABSTRACT FROM AUTHOR]

We developed a wearable swallowing assessment device using a hetero-core fiber-optic pressure sensor for the detection of laryngeal movement during swallowing. The proposed pressure sensor (comfortably attached to the skin of the neck) demonstrated a high sensitivity of 0.592 dB/kPa and a linearity of R2 = 0.995 within a 14 kPa pressure band, which is a suitable pressure for the detection of laryngeal movement. In addition, since the fabricated hetero-core fiber-optic pressure sensor maintains appreciable sensitivity over the surface of the sensor, the proposed wearable swallowing assessment device can accurately track the subtle pressure changes induced by laryngeal movements during the swallowing process. Sixteen male subjects and one female subject were evaluated in a variety of age groups ranging from 30 to 60 years old. For all subjects, characteristic swallowing waveforms (with two valleys based on laryngeal movements consisting of upward, forward, backward, and downward displacements) were acquired using the proposed wearable swallowing assessment device. Since the denoted time of the first valley in the acquired waveform determines the "aging effect", significant differences in swallowing functions among the different age groups were ultimately determined based on the time of the first valley. Additionally, by analyzing each age group using the proposed device, due to p-values being consistently less than 0.05, swallowing times were found to exhibit statistically significant differences within the same groups. [ABSTRACT FROM AUTHOR]

The foot is a vital organ, as it stabilizes the impact forces between the human skeletal system and the ground. Hence, precise foot dimensions are essential not only for custom footwear design, but also for the clinical treatment of foot health. Most existing research on measuring foot dimensions depends on a heavy setup environment, which is costly and ineffective for daily use. In addition, there are several smartphone applications online, but they are not suitable for measuring the exact foot shape for custom footwear, both in clinical practice and public use. In this study, we designed and implemented computer-vision-based smartphone application OptiFit that provides the functionality to automatically measure the four essential dimensions (length, width, arch height, and instep girth) of a human foot from images and 3D scans. We present an instep girth measurement algorithm, and we used a pixel per metric algorithm for measurement; these algorithms were accordingly integrated with the application. Afterwards, we evaluated our application using 19 medical-grade silicon foot models (12 males and 7 females) from different age groups. Our experimental evaluation shows that OptiFit could measure the length, width, arch height, and instep girth with an accuracy of 95.23%, 96.54%, 89.14%, and 99.52%, respectively. A two-tailed paired t-test was conducted, and only the instep girth dimension showed a significant discrepancy between the manual measurement (MM) and the application-based measurement (AM). We developed a linear regression model to adjust the error. Further, we performed comparative analysis demonstrating that there were no significant errors between MM and AM, and the application offers satisfactory performance as a foot-measuring application. Unlike other applications, the iOS application we developed, OptiFit, fulfils the requirements to automatically measure the exact foot dimensions for individually fitted footwear. Therefore, the application can facilitate proper foot measurement and enhance awareness to prevent foot-related problems caused by inappropriate footwear. [ABSTRACT FROM AUTHOR]

The giant panda (Ailuropoda melanoleuca) has long attracted the attention of conservationists as a flagship and umbrella species. Collecting attribute information on the age structure and sex ratio of the wild giant panda populations can support our understanding of their status and the design of more effective conservation schemes. In view of the shortcomings of traditional methods, which cannot automatically recognize the age and sex of giant pandas, we designed a SENet (Squeeze-and-Excitation Network)-based model to automatically recognize the attributes of giant pandas from their vocalizations. We focused on the recognition of age groups (juvenile and adult) and sex of giant pandas. The reason for using vocalizations is that among the modes of animal communication, sound has the advantages of long transmission distances, strong penetrating power, and rich information. We collected a dataset of calls from 28 captive giant panda individuals, with a total duration of 1298.02 s of recordings. We used MFCC (Mel-frequency Cepstral Coefficients), which is an acoustic feature, as inputs for the SENet. Considering that small datasets are not conducive to convergence in the training process, we increased the size of the training data via SpecAugment. In addition, we used focal loss to reduce the impact of data imbalance. Our results showed that the F1 scores of our method for recognizing age group and sex reached 96.46% ± 5.71% and 85.85% ± 7.99%, respectively, demonstrating that the automatic recognition of giant panda attributes based on their vocalizations is feasible and effective. This more convenient, quick, timesaving, and laborsaving attribute recognition method can be used in the investigation of wild giant pandas in the future. [ABSTRACT FROM AUTHOR]

Changes in articular surfaces can be associated with the aging process and as such may lead to quantitative and qualitative impairment of joint motion. This study is aiming to evaluate the age-related quality of the knee joint arthrokinematic motion using nonlinear parameters of the vibroarthrographic (VAG) signal. To analyse the age-related quality of the patellofemoral joint (PFJ), motion vibroarthrography was used. The data that were subject to analysis represent 220 participants divided into five age groups. The VAG signals were acquired during flexion/extension knee motion and described with the following nonlinear parameters: recurrence rate ( R R ) and multi-scale entropy ( M S E ). R R and M S E decrease almost in a linear way with age (main effects of group p < 0.001 ; means ( S D ): R R = 0.101 (0.057) − 0.020 (0.017) ; and M S E = 20.9 (8.56) − 13.6 (6.24)) . The R R post-hoc analysis showed that there were statistically significant differences ( p < 0.01 ) in all comparisons with the exception of the 5th–6th life decade. For M S E , statistically significant differences ( p < 0.01 ) occurred for: 3rd–7th, 4th–7th, 5th–7th and 6th life decades. Our results imply that degenerative age-related changes are associated with lower repeatability, greater heterogeneity in state space dynamics, and greater regularity in the time domain of VAG signal. In comparison with linear VAG measures, our results provide additional information about the nature of changes of the vibration dynamics of PFJ motion with age. [ABSTRACT FROM AUTHOR]

The task to develop a mechanism for predicting the hemodynamic parameters values based on non-invasive hydrocuff technology of a pulse wave signal fixation is described in this study. The advantages and disadvantages of existing methods of recording the ripple curve are noted in the published materials. This study proposes a new hydrocuff method for hemodynamic parameters and blood pressure values measuring. A block diagram of the device being developed is presented. Algorithms for processing the pulse wave contour are presented. A neural network applying necessity for the multiparametric feature space formation is substantiated. The pulse wave contours obtained using hydrocuff technology of oscillation formation for various age groups are presented. According to preliminary estimates, by the moment of the dicrotic surge formation, it is possible to judge the ratio of the heart and blood vessels work, which makes it possible to form an expanded feature space of significant parameters based on neural network classifiers. This study presents the characteristics accounted for creating a database for training a neural network. [ABSTRACT FROM AUTHOR]

Recurrent joint bleeds and silent bleeds are the most common clinical feature in patients with hemophilia. Every bleed causes an immediate inflammatory response and is the leading cause of chronic crippling arthropathy. With the help of infrared thermography we wanted to detect early differences between a group of clinical non-symptomatic children with hemophilia (CWH) with no history of clinically detected joint bleeds and a healthy age-matched group of children. This could help to discover early inflammation and help implement early treatment and preventative strategies. It could be demonstrated that infrared thermography is sensitive enough to detect more signs of early inflammatory response in the CWH than in healthy children. It seems to detect more side differences in temperature than clinical examination of silent symptoms detects tender points. Silent symptoms/tender points seem to be combined with early local inflammation. Using such a non-invasive and sensor-based early detection, prevention of overloading and bleeding might be achieved. [ABSTRACT FROM AUTHOR]

Spirometers are important devices for following up patients with respiratory diseases. These are mainly located only at hospitals, with all the disadvantages that this can entail. This limits their use and consequently, the supervision of patients. Research efforts focus on providing digital alternatives to spirometers. Although less accurate, the authors claim they are cheaper and usable by many more people worldwide at any given time and place. In order to further popularize the use of spirometers even more, we are interested in also providing user-friendly lung-capacity metrics instead of the traditional-spirometry ones. The main objective, which is also the main contribution of this research, is to obtain a person's lung age by analyzing the properties of their exhalation by means of a machine-learning method. To perform this study, 188 samples of blowing sounds were used. These were taken from 91 males (48.4%) and 97 females (51.6%) aged between 17 and 67. A total of 42 spirometer and frequency-like features, including gender, were used. Traditional machine-learning algorithms used in voice recognition applied to the most significant features were used. We found that the best classification algorithm was the Quadratic Linear Discriminant algorithm when no distinction was made between gender. By splitting the corpus into age groups of 5 consecutive years, accuracy, sensitivity and specificity of, respectively, 94.69%, 94.45% and 99.45% were found. Features in the audio of users' expiration that allowed them to be classified by their corresponding lung age group of 5 years were successfully detected. Our methodology can become a reliable tool for use with mobile devices to detect lung abnormalities or diseases. [ABSTRACT FROM AUTHOR]

Social interactions significantly impact the quality of life for people with special needs (e.g., older adults with dementia and children with autism). They may suffer loneliness and social isolation more often than people without disabilities. There is a growing demand for technologies to satisfy the social needs of such user groups. However, evaluating these systems can be challenging due to the extra difficulty of gathering data from people with special needs (e.g., communication barriers involving older adults with dementia and children with autism). Thus, in this systematic review, we focus on studying data gathering methods for evaluating socially assistive systems (SAS). Six academic databases (i.e., Scopus, Web of Science, ACM, Science Direct, PubMed, and IEEE Xplore) were searched, covering articles published from January 2000 to July 2021. A total of 65 articles met the inclusion criteria for this systematic review. The results showed that existing SASs most often targeted people with visual impairments, older adults, and children with autism. For instance, a common type of SASs aimed to help blind people perceive social signals (e.g., facial expressions). SASs were most commonly assessed with interviews, questionnaires, and observation data. Around half of the interview studies only involved target users, while the other half also included secondary users or stakeholders. Questionnaires were mostly used with older adults and people with visual impairments to measure their social interaction, emotional state, and system usability. A great majority of observational studies were carried out with users in special age groups, especially older adults and children with autism. We thereby contribute an overview of how different data gathering methods were used with various target users of SASs. Relevant insights are extracted to inform future development and research. [ABSTRACT FROM AUTHOR]

Smart home services (SHS) should support the positive experiences of the elderly in homes with a focus on getting closer to nature. The study identified the services preferred by the elderly through a survey on the biophilic experience-based SHS, and to discuss the configuration of the sensors and devices required to provide the service. We reorganized the biophilic experience-based SHS and related sensors and devices, focusing on our previous study, and developed a survey instrument. A preference survey was conducted on 250 adults aged 20 and older, and the SPSS program was used for a factor analysis and independent two-sample T-test. We derived six factors for biophilic experience-based SHS. Compared to other age groups, the elderly preferred services that were mainly attributed to factors such as 'Immersion and interaction with nature' (A), 'Management of well-being and indoor environmental quality (IEQ)' (B), and 'Natural process and systems' (F). We proposed 15 prioritized services, along with their sensor and device configurations, in consideration of service provision regarding the elderly's preferences and universality. This study contributes to new developments in elderly-friendly smart home research by converting bio-friendly ideas into the market in the development of medical services and SHS for the elderly. [ABSTRACT FROM AUTHOR]

Quality of life (QoL) is an important parameter that affects the choice of therapy. Assessment of QoL and satisfaction with therapy using the rtCGM in children with T1D aged < 7 years was conducted. The study group consisted of 38 children with T1D aged < 7 years (34% aged 2–4, 66% aged 5–7 years), HbA1c: 6.53 ± 0.63%, duration of diabetes: 2.6 ± 1.6 years, treated with an rtCGM-augmented insulin pump for 1.92 ± 1.15 years. Two anonymous surveys were conducted: a. PedsQL3.0 diabetes standardized questionnaire—QoL assessment among age groups: 2–4/5–7 years. b. An original survey assessing the CGM use satisfaction. The mean scores in PedsQL3.0: communication 75%, worries 30%, treatment 70%, and problems associated with diabetes 65%. The QoL scale is: 0–19% very low, 20–39% low, 40–59% moderate, 60–79% high, 80–100% very high. The most frequently reported concerns were long-term diabetes complications and prick pain. Satisfaction with CGM use was high (68% in group aged 5–7 and 92% 2–4 years). Twenty-seven (71%) caregivers confirmed the positive effect of CGM on sleep. During the use of rtCGM a high quality of life was reported, and the quality of sleep in their caregivers was increased. [ABSTRACT FROM AUTHOR]

Accelerometer-based fitness trackers and smartwatches are proliferating with incessant attention towards health tracking. Despite their growing popularity, accurately measuring hallmark measures of physical activities has yet to be accomplished in adults of all ages. In this work, we evaluated the performance of four machine learning models: decision tree, random forest, extreme gradient boosting (XGBoost) and least absolute shrinkage and selection operator (LASSO), to estimate the hallmark measures of physical activities in young (20–50 years), middle-aged (50–70 years], and older adults (70–89 years]. Our models were built to recognize physical activity types, recognize physical activity intensities, estimate energy expenditure (EE) and recognize individual physical activities using wrist-worn tri-axial accelerometer data (33 activities per participant) from a large sample of participants (n = 253, 62% women, aged 20–89 years old). Results showed that the machine learning models were quite accurate at recognizing physical activity type and intensity and estimating energy expenditure. However, models performed less optimally when recognizing individual physical activities. F1-Scores derived from XGBoost's models were high for sedentary (0.955–0.973), locomotion (0.942–0.964) and lifestyle (0.913–0.949) activity types with no apparent difference across age groups. Low (0.919–0.947), light (0.813–0.828) and moderate (0.846–0.875) physical activity intensities were also recognized accurately. The root mean square error range for EE was approximately 1 equivalent of resting EE [0.835–1.009 METs]. Generally, random forest and XGBoost models outperformed other models. In conclusion, machine learning models to label physical activity types, activity intensity and energy expenditure are accurate and there are minimal differences in their performance across young, middle-aged and older adults. [ABSTRACT FROM AUTHOR]

We present a multi-task learning-based convolutional neural network (MTL-CNN) able to estimate multiple tags describing face images simultaneously. In total, the model is able to estimate up to 74 different face attributes belonging to three distinct recognition tasks: age group, gender and visual attributes (such as hair color, face shape and the presence of makeup). The proposed model shares all the CNN’s parameters among tasks and deals with task-specific estimation through the introduction of two components: (i) a gating mechanism to control activations’ sharing and to adaptively route them across different face attributes; (ii) a module to post-process the predictions in order to take into account the correlation among face attributes. The model is trained by fusing multiple databases for increasing the number of face attributes that can be estimated and using a center loss for disentangling representations among face attributes in the embedding space. Extensive experiments validate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]