Wearable Camera Research Articles

User Experience and Compliance Rates of Remote Monitoring Technologies in Alzheimer’s Disease: Preliminary Results of the RADAR‐AD Study

AbstractBackgroundUnlike traditional pen‐and‐paper clinical assessments, remote monitoring technologies (RMTs) can assess participants’ function continuously and objectively during activities of daily living. Despite the potential of RMTs to assess function or assist early disease detection, there is skepticism that age or impairment may render participants unable or unwilling to comply with complex RMT protocols. The 8‐week RADAR‐AD study protocol asks Alzheimer’s disease (AD) patients in all disease stages to wear, amongst others, two activity trackers (Axivity, Fitbit), a wearable camera (Autographer), and use an active smartphone app (Mezurio), with a daily schedule totaling over 400 separate tasks and a parallel schedule for caregivers. This abstract presents preliminary compliance data and participant feedback.MethodWe included amyloid negative control participants (n=47) and amyloid positive preclinical (n=16), prodromal (n=19) and mild‐to‐moderate (n=16) AD participants (Table 1), based on MMSE and CDR scores, in 8 European countries. Metrics based on device data were calculated for participant commitment (study duration completion percentage or wear time) and compliance (task completion percentage). Patient experiences were monitored using bi‐weekly semi‐structured phone interviews.ResultMezurio average commitment and compliance were 90% and 85% respectively (Table 2) and were comparable across diagnostic groups. Wear time of Axivity, Fitbit and wearable camera was high (>86%), with 48% of those offered adopting the optional wearable camera (Table 2). Participant feedback was positive, with relatively few problems reported (Figure 1).ConclusionRADAR‐AD shows that age and impairment are not necessarily barriers to successful deployment of complex RMT study protocols. We acknowledge that our high commitment and compliance rates benefitted from an active rapport between researchers, participants and care givers, which may not be reproducible in every scenario. Nevertheless, this study shows that AD patients are able and willing to comply and engage with technology.This work has received support from the EU/EFPIA Innovative Medicines Initiative Joint Undertaking (grant No 806999). www.imi.europa.eu. This communication reflects the views of the RADAR‐AD consortium and neither IMI nor the European Union and EFPIA are liable for any use that may be made of the information contained herein.

P079 A new lens on preschool sleep: Using wearable cameras to capture pre-bedtime activities

Abstract Introduction While bedtime routines are often recommended to address behavioural sleep problems in young children, these recommendations are based on parental subjective assessment of behaviours which may not be accurate. The aim of this study was to establish the feasibility of using wearable cameras to objectively measure bedtime routines and pre-bedtime screen use in preschool-aged children. Methods Sixteen children aged 3-4 years wore a GoPro video camera for two hours prior to bedtime on three consecutive evenings. Subsequent sleep parameters were monitored for seven nights by actigraphy. Standardised questionnaires and a sleep diary provided traditional measures of child sleep behaviours. Feasibility was determined by adherence to protocol and ability to develop a reliable behavioural coding scheme from the video data. Acceptability was determined through feedback questionnaires and exit interviews. Results Participant families found the cameras acceptable, with the majority of parents agreeing the cameras were easy to use (94%) and keep on their child (81%). Parents felt that wearing the camera did not change how they parented their child (88%), or their child’s behaviours (81%). A behavioural coding scheme incorporating self-care activities, nutrition, reading, physical contact and screen use (time, device type, media content) was successfully developed with interrater reliability established by two independent coders (kappa >0.8). Conclusions Wearable cameras provide a feasible and acceptable method to measure bedtime routines and screen use in preschool-aged children. This appears to be is the first study to use wearable cameras to objectively measure pre-bedtime activities in this age-group.

GO-Finder: A Registration-free Wearable System for Assisting Users in Finding Lost Hand-held Objects

People spend an enormous amount of time and effort looking for lost objects. To help remind people of the location of lost objects, various computational systems that provide information on their locations have been developed. However, prior systems for assisting people in finding objects require users to register the target objects in advance. This requirement imposes a cumbersome burden on the users, and the system cannot help remind them of unexpectedly lost objects. We propose GO-Finder (“Generic Object Finder”), a registration-free wearable camera-based system for assisting people in finding an arbitrary number of objects based on two key features: automatic discovery of hand-held objects and image-based candidate selection. Given a video taken from a wearable camera, GO-Finder automatically detects and groups hand-held objects to form a visual timeline of the objects. Users can retrieve the last appearance of the object by browsing the timeline through a smartphone app. We conducted user studies to investigate how users benefit from using GO-Finder. In the first study, we asked participants to perform an object retrieval task and confirmed improved accuracy and reduced mental load in the object search task by providing clear visual cues on object locations. In the second study, the system’s usability on a longer and more realistic scenario was verified, accompanied by an additional feature of context-based candidate filtering. Participant feedback suggested the usefulness of GO-Finder also in realistic scenarios where more than one hundred objects appear.

Assessing the association between urban features and human physiological stress response using wearable sensors in different urban contexts.

Public open space (POS) plays a significant role in fostering human health and wellbeing in cities. A major limitation of current research on POS and health is that there is little attention on the role of various urban features on people's mental health, in different urban context. This study employed wearable sensors (a wearable camera, Empatica 4 wristband and a GPS device) to measure human physiological responses to urban indicators, objectively. To do this, we selected six kinds of public open space (water area, transit area, green area, commercial area, motor traffic area and mixed office and residential area) and recruited 86 participants for an experimental study. Next, we detected urban features by using Microsoft Cognitive Services (MCS) and calculated a change score to assess human physiological stress responses based on galvanic skin response (GSR) and skin temperature from the wristband. Lastly, we applied random effect model and geographically weighted regression analysis to examine the relationship between urban indicators and human physiological stress responses. The findings show that urban flow (vehicles, bikes and people), waterbodies, greenery and places to sit are associated with the changes of human physiological stress response. The findings indicate that the type of urban context may confound the effect of green and blue urban features; i.e., the effect on physiological stress response can be positive or negative depending on the context. The paper highlights the relevance of considering urban context in research on associations between urban features and stress response.

Measuring the Usability of eHealth Solutions for Patients With Parkinson Disease: Observational Study.

Parkinson disease (PD) is a neurodegenerative disorder with a variety of motor and nonmotor symptoms. Many of these symptoms can be monitored by eHealth solutions, including smartphone apps, wearable sensors, and camera systems. The usability of such systems is a key factor in long-term use, but not much is known about the predictors of successful use and preferable methods to assess usability in patients with PD. This study tested methods to assess usability and determined prerequisites for successful use in patients with PD. We performed comprehensive usability assessments with 18 patients with PD using a mixed methods usability battery containing the System Usability Scale, a rater-based evaluation of device-specific tasks, and qualitative interviews. Each patient performed the usability battery with 2 of 3 randomly assigned devices: a tablet app, wearable sensors, and a camera system. The usability battery was administered at the beginning and at the end of a 4-day testing period. Between usability batteries, the systems were used by the patients during 3 sessions of motor assessments (wearable sensors and camera system) and at the movement disorder ward (tablet app). In this study, the rater-based evaluation of tasks discriminated the best between the 3 eHealth solutions, whereas subjective modalities such as the System Usability Scale were not able to distinguish between the systems. Successful use was associated with different clinical characteristics for each system: eHealth literacy and cognitive function predicted successful use of the tablet app, and better motor function and lower age correlated with the independent use of the camera system. The successful use of the wearable sensors was independent of clinical characteristics. Unfortunately, patients who were not able to use the devices well provided few improvement suggestions in qualitative interviews. eHealth solutions should be developed with a specific set of patients in mind and subsequently tested in this cohort. For a complete picture, usability assessments should include a rater-based evaluation of task performance, and there is a need to develop strategies to circumvent the underrepresentation of poorly performing patients in qualitative usability research.

The Feasibility of Integrating Wearable Cameras and Health Trackers for Measuring Personal Exposure to Urban Features

Built environment factors such as greenery, walkability, and crowd density are related to physical activity and mental health. New emerging wearable sensors provide an opportunity to objectively monitor human exposure to street-level urban features. However, very few studies have demonstrated how to objectively measure the association between the built environment, human emotions, and health. This pilot study proposes a new approach that employs a FrontRow wearable lifestyle camera, a GPS tracker, and an Empatica 4 wristband as a sensor package to track individuals during their everyday activities. Machine-learning methods are adopted to extract urban features. For this study, volunteers were asked to conduct a self-led city tour in Roskilde, Denmark, while using the wearable sensors. Study results demonstrate the feasibility of the proposed approach and the potential for using integrated, multi-sourced data in the study of urban health.

Visual Object Tracking in First Person Vision

The understanding of human-object interactions is fundamental in First Person Vision (FPV). Visual tracking algorithms which follow the objects manipulated by the camera wearer can provide useful information to effectively model such interactions. In the last years, the computer vision community has significantly improved the performance of tracking algorithms for a large variety of target objects and scenarios. Despite a few previous attempts to exploit trackers in the FPV domain, a methodical analysis of the performance of state-of-the-art trackers is still missing. This research gap raises the question of whether current solutions can be used “off-the-shelf” or more domain-specific investigations should be carried out. This paper aims to provide answers to such questions. We present the first systematic investigation of single object tracking in FPV. Our study extensively analyses the performance of 42 algorithms including generic object trackers and baseline FPV-specific trackers. The analysis is carried out by focusing on different aspects of the FPV setting, introducing new performance measures, and in relation to FPV-specific tasks. The study is made possible through the introduction of TREK-150, a novel benchmark dataset composed of 150 densely annotated video sequences. Our results show that object tracking in FPV poses new challenges to current visual trackers. We highlight the factors causing such behavior and point out possible research directions. Despite their difficulties, we prove that trackers bring benefits to FPV downstream tasks requiring short-term object tracking. We expect that generic object tracking will gain popularity in FPV as new and FPV-specific methodologies are investigated.

Representational dynamics of memories for real-life events

The continuous flow of experience that characterizes real-life events is not recorded as such in episodic memory but is condensed as a succession of event segments separated by temporal discontinuities. To unravel the neural basis of this representational structure, we recorded real-life events using wearable camera technology and used fMRI to investigate brain activity during their temporal unfolding in memory. We found that, compared to the representation of static scenes in memory, dynamically unfolding memory representations were associated with greater activation of the posterior medial episodic network. Strikingly, by analyzing the autocorrelation of brain activity patterns at successive time points throughout the retrieval period, we found that this network showed higher temporal dynamics when recalling events that included a higher density of event segments. These results reveal the key role of the posterior medial network in representing the dynamic unfolding of the event segments that constitute real-world memories.

Wearable Cameras Reveal Large Intra-Individual Variability in Timing of Eating among Young Adults.

Studies have shown that young adults follow less structured eating patterns compared with older cohorts. This may have implications for dietary assessment methods which rely on memory and structured meal patterns. Our aim was to describe the intra-individual variation of eating times in young adults aged 18–30 years. Participants (n = 41) wore an Autographer camera that captured first-person perspective images every 30 s for three consecutive days. All images were timestamped and those showing food consumption were used to extract data such as the timing of the first and last eating occasions (EOs), number of EOs per day, and length of eating window. Intra-individual variability was calculated from these data using composite phase deviation (CPD) and coefficient of variation (CV). The number of individuals with high or very high variability was 28 and 18 for timing of first and last EOs, respectively (CPD > 1.70), and 27 and 17 for number of EOs and eating window, respectively (CV > 20%). In this sample of young adults, the lack of regularity in eating patterns should be considered when selecting a dietary assessment method.

Perspectives and expectations of stroke survivors using egocentric cameras for monitoring hand function at home: a mixed methods study

Purpose Most stroke survivors have remaining upper limb impairment six months after stroke and require additional rehabilitation and help from family members to enhance their performance of daily activities. First-person (egocentric) video has been proposed to capture the activities of daily living (ADLs) of stroke survivors in order to assess their hand function at home. This study explored the experiences and expectations of stroke survivors regarding the use of egocentric cameras in daily life for rehabilitation applications. Methods Twenty-one chronic stroke survivors recruited for the study were asked to record three sessions of 1.5 h of video of their ADLs at home over two weeks. Their experiences and expectations after completing the recordings were discussed using a structured questionnaire and a semi-structured interview. The questionnaire and interview data were analysed using descriptive statistics and content analysis, respectively. The results were further integrated using a mixed methods analysis for mutual explanation and elaboration. Results The themes generated were Camera Usability, Privacy Concerns Related to Home Recordings, Future Use of the Camera in Public, and Information Usefulness. The participants perceived that the camera was easy to use, the information obtained from the recordings was beneficial, and no major concerns about recording at home. A discreet camera and a solution to privacy issues were prerequisites to recording tasks in public. Conclusions There was high acceptance among stroke survivors regarding the use of wearable cameras for rehabilitation purposes in the future. Concerns to be managed include discomfort, self-consciousness, and the privacy of others. Implications for rehabilitation The egocentric camera was easy for the stroke survivors to use at home. However, they expressed a preference for cameras to be less noticeable and lighter in the future to minimize self-consciousness and discomfort. Expectations for future use of an egocentric camera for upper limb rehabilitation at home from the perspectives of stroke survivors included receiving feedback on their hand function in daily life and guidance on how to improve function. Privacy concerns of stroke survivors regarding recording activities of daily living were mostly avoidable by planning in advance. However, some personal hygiene tasks and virtual meetings were recorded by accident. A checklist of common activities that may raise privacy issues can be provided along with the camera to serve as a reminder to avoid these issues.

Energy intake estimation using a novel wearable sensor and food images in a laboratory (pseudo-free-living) meal setting: quantification and contribution of sources of error.

Dietary assessment methods not relying on self-report are needed. The Automatic Ingestion Monitor 2 (AIM-2) combines a wearable camera that captures food images with sensors that detect food intake. We compared energy intake (EI) estimates of meals derived from AIM-2 chewing sensor signals, AIM-2 images, and an internet-based diet diary, with researcher conducted weighed food records (WFR) as the gold standard. Thirty adults wore the AIM-2 for meals self-selected from a university food court on one day in mixed laboratory and free-living conditions. Daily EI was determined from a sensor regression model, manual image analysis, and a diet diary and compared with that from WFR. A posteriori analysis identified sources of error for image analysis and WFR differences. Sensor-derived EI from regression modeling (R2 = 0.331) showed the closest agreement with EI from WFR, followed by diet diary estimates. EI from image analysis differed significantly from that by WFR. Bland-Altman analysis showed wide limits of agreement for all three test methods with WFR, with the sensor method overestimating at lower and underestimating at higher EI. Nutritionist error in portion size estimation and irreconcilable differences in portion size between food and nutrient databases used for WFR and image analyses were the greatest contributors to image analysis and WFR differences (44.4% and 44.8% of WFR EI, respectively). Estimation of daily EI from meals using sensor-derived features offers a promising alternative to overcome limitations of self-report. Image analysis may benefit from computerized analytical procedures to reduce identified sources of error.

Alaskan children’s perspectives of environmental stewardship in a changing Arctic environment

ABSTRACT This chapter magnifies the voices of Alaskan children in considering education for sustainability in the Arctic. The collaborative research explored how children understood and enacted environmental stewardship in three distinct Alaskan locations: Interior Alaska, the Kenai peninsula, and a rural southwest Alaska Native village. Honoring children’s agency, the study involved child-centered research methods, including children’s drawings and descriptions, role-playing, class discussions, and video tours utilizing wearable cameras. Findings revealed common themes of environmental stewardship, yet the way children perceived and enacted stewardship varied according to the social, cultural, and geographical contexts. Cleaning up litter was perceived as immediate and important for children from Kenai and the southwest Alaskan village, yet it was scarcely mentioned by Interior Alaskan children. Interior Alaskan children emphasized pet care, while children from Kenai and the southwest Alaskan village discussed animal care in relation to hunting and fishing ethics. Care for plants was less common than care for snow. Children’s spatial autonomy, sense of belonging and personal connection with place, plays an important role in the development of competencies to live more sustainably on the land. Findings point towards contextualized and child-centered approaches to promote children’s agency to act in and for their environments.

Egocentric Human Trajectory Forecasting With a Wearable Camera and Multi-Modal Fusion

In this paper, we address the problem of forecasting the trajectory of an egocentric camera wearer (ego-person) in crowded spaces. The trajectory forecasting ability learned from the data of different camera wearers walking around in the real world can be transferred to assist visually impaired people in navigation, as well as to instill human navigation behaviours in mobile robots, enabling better human-robot interactions. To this end, a novel egocentric human trajectory forecasting dataset was constructed, containing real trajectories of people navigating in crowded spaces wearing a camera, as well as extracted rich contextual data. We extract and utilize three different modalities to forecast the trajectory of the camera wearer, i.e., his/her past trajectory, the past trajectories of nearby people, and the environment such as the scene semantics or the depth of the scene. A Transformer-based encoder-decoder neural network model, integrated with a novel cascaded cross-attention mechanism that fuses multiple modalities, has been designed to predict the future trajectory of the camera wearer. Extensive experiments have been conducted, with results showing that our model outperforms the state-of-the-art methods in egocentric human trajectory forecasting.

APPLE MOTS: Detection, Segmentation and Tracking of Homogeneous Objects Using MOTS

Current multi object tracking and segmentation (MOTS) methods made great progress for the simultaneous detection and tracking of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">heterogeneous</i> objects like cars and pedestrians. Nevertheless, all of these scenes consisted of dissimilar objects, which are easier to track than homogeneous and smaller objects, as those are more similar in appearance. Therefore, this is the first paper that explores the implementation of MOTS algorithms for the simultaneous detection and tracking of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">homogeneous</i> objects. Towards this end, video data was acquired in an apple orchard using a wearable camera and unmanned aerial vehicles (UAV). The dataset, called APPLE MOTS, contains almost 86000 manually annotated apple masks and is the first public dataset in which apple instances are temporally consistent labelled across frames. Implementation of the MOTS architectures called TrackR-CNN and PointTrack indicates that they could be suitable for the joint detection (MOTSP: 80.4) and tracking (sMOTSA: 38.7, MOTSA: 52.9) of apples. This letter exposes the challenge of tracking homogeneous objects due to their similar shape and colour while detection performance remains state-of-the-art.

SmartAct: Energy Efficient and Real-Time Hand-to-Mouth Gesture Detection Using Wearable RGB-T.

Researchers have been leveraging wearable cameras to both visually confirm and automatically detect individuals' eating habits. However, energy-intensive tasks such as continuously collecting and storing RGB images in memory, or running algorithms in real-time to automate detection of eating, greatly impacts battery life. Since eating moments are spread sparsely throughout the day, battery life can be mitigated by recording and processing data only when there is a high likelihood of eating. We present a framework comprising a golf-ball sized wearable device using a low-powered thermal sensor array and real-time activation algorithm that activates high-energy tasks when a hand-to-mouth gesture is confirmed by the thermal sensor array. The high-energy tasks tested are turning on the RGB camera (Trigger RGB mode) and running inference on an on-device machine learning model (Trigger ML mode). Our experimental setup involved the design of a wearable camera, 6 participants collecting 18 hours of data with and without eating, the implementation of a feeding gesture detection algorithm on-device, and measures of power saving using our activation method. Our activation algorithm demonstrates an average of at-least 31.5% increase in battery life time, with minimal drop of recall (5%) and without impacting the accuracy of detecting eating (a slight 4.1% increase in F1-Score).

A multimodal hyperlapse method based on video and songs’ emotion alignment

With the recent growth in the use of social media and new digital devices like smartphones and wearable cameras, people are often recording long first-person videos of their daily activities. These videos are usually very long and tiring to watch, bringing the need to speed them up. Recent fast-forward methods do not consider the background music to be inserted into the video, which could make them even more enjoyable. In this paper, we present a new fast-forward method that considers the information present in the video and the background music. We use neural networks to automatically recognize the emotions induced in the video and song and combine the contents in the accelerated video through a new method of frame selection that has as main objective to maximize the similarity of the induced emotions. We present quantitative and qualitative experiments on a large dataset with different videos and songs, showing that our method achieves the best performance in matching emotion similarity, also keeping the video’s visual quality.

Comparison Of Wearable Camera Image Physical Activity Behavior Estimates To Direct Observation

PURPOSE: To assess the accuracy and precision of physical activity (PA) behavior estimates from wearable camera (WC) still-images (IMGs) to video-recorded direct observation (DO). METHODS: Participants (n = 19) wore a WC above the xiphoid process while simultaneously being video recorded by researchers, during 3 visits, in a place of the participants choice. WC was set to the highest capture rate. A PA behavior annotation protocol was developed from the Compendium of Physical Activities. Videos and IMGs were annotated using Noldus Observer XT14 and Image Browser software, respectively. For analysis, PA behavior annotations were collapsed into 8 domains. A linear mixed effects model was used to calculate statistical bias and 95% CIs of IMGs to DO behavior times to determine accuracy and precision. Confusion matrices were computed to determine trends in behavior misclassification by IMGs against DO. One-way random effects interclass correlation coefficient (ICC) was calculated between domains to assess reliability. RESULTS: 53 visits were analyzed with a total time of 4381.3 min (mean 82.7 ± 19.2 min). Fig. 1 shows bias and 95% confidence intervals between IMGs and DO for each behavior domain. Discrepancies of PA behavior min estimation by IMGs against DO annotations were statistically insignificant at p > 0.05. Household had the smallest bias (0.09 ± 6.2 min) whereas, sport & exercise had the largest (1.61 ± 9.32 min). From confusion matrices, the domain “other” had the largest misclassification, as IMG annotation scored such events as “household” 26.3% of the time. ICC between IMGs and DO was excellent (ICC = 0.8945 [0.8938, 0.8953]). CONCLUSIONS: WC IMG annotation is accurate and precise when estimating PA behavior compared to annotated, DO videos. Future work examining WC use in more heterogeneous populations and over longer periods of time would further advance the application of this technology in measurement research.

Assessing Posture And Activity Intensity Agreement Between Wearable Camera Images And Criterion Direct Observation

PURPOSE: To compare posture and activity intensity estimates annotated from wearable camera images (WCI) to criterion video-recorded direct observation (DO). METHODS: 19 healthy adults were observed over 53 visits (mean visit length = 89.7 ± 19.2 minutes), in a location of their choice. During these visits, participants were video-recorded performing free-living activities of daily living, while wearing a wearable camera at chest level. The camera was set to a high capture rate, where it took first-person images. Researchers annotated the WCI and DO videos for posture (sitting, standing, and movement). Activity intensity (sedentary, light, moderate-vigorous) from DO videos was determined from posture, while WCI were coded using a schema with matching Taylor codes to the Compendium of Physical Activities to extrapolate activity intensity. For analysis, a linear mixed-effects model calculated statistical bias and 95% CIs of WCI posture and activity intensity time in minutes to DO. Cohen’s weighted Kappa statistic (Κ) was calculated to assess the agreement of posture and activity intensity annotations to criterion DO. RESULTS: 4754 minutes of WC and DO data were collected. Significant bias was calculated for all outcome measures except movement posture (Figure 1). Kappa showed substantial agreement between WCI and DO to assess posture (K = 0.76), and moderate agreement between WCI and DO for estimating intensity (K = 0.55). CONCLUSIONS: When assessing posture in healthy adults, WCI show substantial agreement with criterion DO. WCI show moderate agreement with DO when used to estimate activity intensity. Future studies should explore WCI’s estimates of posture and activity intensity in more diverse populations and validate WCI activity intensity estimations against gold standard measurements of energy expenditure.

Investigating Wrist-Based Acceleration Summary Measures across Different Sample Rates towards 24-Hour Physical Activity and Sleep Profile Assessment.

Wrist-worn wearable sensors have attracted considerable research interest because of their potential in providing continuous, longitudinal, non-invasive measurements, leading to insights into Physical Activity (PA), sleep, and circadian variability. Three key practical considerations for research-grade wearables are as follows: (a) choosing an appropriate sample rate, (b) summarizing raw three-dimensional accelerometry data for further processing (accelerometry summary measures), and (c) accurately estimating PA levels and sleep towards understanding participants’ 24-hour profiles. We used the CAPTURE-24 dataset, where 148 participants concurrently wore a wrist-worn three-dimensional accelerometer and a wearable camera over approximately 24 h to obtain minute-by-minute labels: sleep; and sedentary light, moderate, and vigorous PA. We propose a new acceleration summary measure, the Rate of Change Acceleration Movement (ROCAM), and compare its performance against three established approaches summarizing three-dimensional acceleration data towards replicating the minute-by-minute labels. Moreover, we compare findings where the acceleration data was sampled at 10, 25, 50, and 100 Hz. We demonstrate the competitive advantage of ROCAM towards estimating the five labels (80.2% accuracy) and building 24-hour profiles where the sample rate of 10 Hz is fully sufficient. Collectively, these findings provide insights facilitating the deployment of large-scale longitudinal actigraphy data processing towards 24-hour PA and sleep-profile assessment.

Visual Loop Closure Detection With Instance Segmentation and Image Inpainting in Dynamic Scenes Using Wearable Camera

Dynamic interference is an important factor that affects the accuracy of loop closure detection in visual SLAM. Traditional methods that remove the feature points in dynamic regions are prone to insufficient image features, which will decrease the reliability of loop closure detection. To overcome this problem, a novel loop closure detection approach based on image inpainting and feature selection is proposed. A dynamic object instance segmentation guided image inpainting network is utilized to segment and repair dynamic regions in scene images. To avoid the negative influence of image inpainting, an image quality evaluation network is used to confine the extraction of Superpoint features only in the properly inpainted patches. Therefore, valid Superpoint features only in the areas with high inpainting qualities are selected as the input of the Bag of Words model for loop-closure detection. We tested the proposed method on both the KITTI dataset and a wearable camera system. Experimental results indicate that the proposed method improves the accuracy of loop closure detection in dynamic scenes.