Corrigendum: Applications and techniques for fast machine learning in science.

[This corrects the article DOI: 10.3389/fdata.2022.787421.].

ActiSight: Wearer Foreground Extraction Using a Practical RGB-Thermal Wearable.

Wearable cameras provide an informative view of wearer activities, context, and interactions. Video obtained from wearable cameras is useful for life-logging, human activity recognition, visual confirmation, and other tasks widely utilized in mobile computing today. Extracting foreground information related to the wearer and separating irrelevant background pixels is the fundamental operation underlying these tasks. However, current wearer foreground extraction methods that depend on image data alone are slow, energy-inefficient, and even inaccurate in some cases, making many tasks-like activity recognition- challenging to implement in the absence of significant computational resources. To fill this gap, we built ActiSight, a wearable RGB-Thermal video camera that uses thermal information to make wearer segmentation practical for body-worn video. Using ActiSight, we collected a total of 59 hours of video from 6 participants, capturing a wide variety of activities in a natural setting. We show that wearer foreground extracted with ActiSight achieves a high dice similarity score while significantly lowering execution time and energy cost when compared with an RGB-only approach.

Impacts of Image Obfuscation on Fine-grained Activity Recognition in Egocentric Video.

Automated detection and validation of fine-grained human activities from egocentric vision has gained increased attention in recent years due to the rich information afforded by RGB images. However, it is not easy to discern how much rich information is necessary to detect the activity of interest reliably. Localization of hands and objects in the image has proven helpful to distinguishing between hand-related fine-grained activities. This paper describes the design of a hand-object-based mask obfuscation method (HOBM) and assesses its effect on automated recognition of fine-grained human activities. HOBM masks all pixels other than the hand and object in-hand, improving the protection of personal user information (PUI). We test a deep learning model trained with and without obfuscation using a public egocentric activity dataset with 86 class labels and achieve almost similar classification accuracies (2% decrease with obfuscation). Our findings show that it is possible to protect PUI at smaller image utility costs (loss of accuracy).

Detecting Screen Presence with Activity-Oriented RGB Camera in Egocentric Videos.

Screen time is associated with several health risk behaviors including mindless eating, sedentary behavior, and decreased academic performance. Screen time behavior is traditionally assessed with self-report measures, which are known to be burdensome, inaccurate, and imprecise. Recent methods to automatically detect screen time are geared more towards detecting television screens from wearable cameras that record high-resolution video. Activity-oriented wearable cameras (i.e., cameras oriented towards the wearer with a fisheye lens) have recently been designed and shown to reduce privacy concerns, yet pose a greater challenge in capturing screens due to their orientation and fewer pixels on target. Methods that detect screens from low-power, low-resolution wearable camera video are needed given the increased adoption of such devices in longitudinal studies. We propose a method that leverages deep learning algorithms and lower-resolution images from an activity-oriented camera to detect screen presence from multiple types of screens with high variability of pixel on target (e.g., near and far TV, smartphones, laptops, and tablets). We test our system in a real-world study comprising 10 individuals, 80 hours of data, and 1.2 million low-resolution RGB frames. Our results outperform existing state-of-the-art video screen detection methods yielding an F1-score of 81%. This paper demonstrates the potential for detecting screen-watching behavior in longitudinal studies using activity-oriented cameras, paving the way for a nuanced understanding of screen time's relationship with health risk behaviors.

Applications and Techniques for Fast Machine Learning in Science.

In this community review report, we discuss applications and techniques for fast machine learning (ML) in science-the concept of integrating powerful ML methods into the real-time experimental data processing loop to accelerate scientific discovery. The material for the report builds on two workshops held by the Fast ML for Science community and covers three main areas: applications for fast ML across a number of scientific domains; techniques for training and implementing performant and resource-efficient ML algorithms; and computing architectures, platforms, and technologies for deploying these algorithms. We also present overlapping challenges across the multiple scientific domains where common solutions can be found. This community report is intended to give plenty of examples and inspiration for scientific discovery through integrated and accelerated ML solutions. This is followed by a high-level overview and organization of technical advances, including an abundance of pointers to source material, which can enable these breakthroughs.

SmokeMon: Unobtrusive Extraction of Smoking Topography Using Wearable Energy-Efficient Thermal.

Smoking is the leading cause of preventable death worldwide. Cigarette smoke includes thousands of chemicals that are harmful and cause tobacco-related diseases. To date, the causality between human exposure to specific compounds and the harmful effects is unknown. A first step in closing the gap in knowledge has been measuring smoking topography, or how the smoker smokes the cigarette (puffs, puff volume, and duration). However, current gold-standard approaches to smoking topography involve expensive, bulky, and obtrusive sensor devices, creating unnatural smoking behavior and preventing their potential for real-time interventions in the wild. Although motion-based wearable sensors and their corresponding machine-learned models have shown promise in unobtrusively tracking smoking gestures, they are notorious for confounding smoking with other similar hand-to-mouth gestures such as eating and drinking. In this paper, we present SmokeMon, a chest-worn thermal-sensing wearable system that can capture spatial, temporal, and thermal information around the wearer and cigarette all day to unobtrusively and passively detect smoking events. We also developed a deep learning-based framework to extract puffs and smoking topography. We evaluate SmokeMon in both controlled and free-living experiments with a total of 19 participants, more than 110 hours of data, and 115 smoking sessions achieving an F1-score of 0.9 for puff detection in the laboratory and 0.8 in the wild. By providing SmokeMon as an open platform, we provide measurement of smoking topography in free-living settings to enable testing of smoking topography in the real world, with potential to facilitate timely smoking cessation interventions.

Continuous Detection of Physiological Stress with Commodity Hardware.

Timely detection of an individual's stress level has the potential to improve stress management, thereby reducing the risk of adverse health consequences that may arise due to mismanagement of stress. Recent advances in wearable sensing have resulted in multiple approaches to detect and monitor stress with varying levels of accuracy. The most accurate methods, however, rely on clinical-grade sensors to measure physiological signals; they are often bulky, custom made, and expensive, hence limiting their adoption by researchers and the general public. In this article, we explore the viability of commercially available off-the-shelf sensors for stress monitoring. The idea is to be able to use cheap, nonclinical sensors to capture physiological signals and make inferences about the wearer's stress level based on that data. We describe a system involving a popular off-the-shelf heart rate monitor, the Polar H7; we evaluated our system with 26 participants in both a controlled lab setting with three well-validated stress-inducing stimuli and in free-living field conditions. Our analysis shows that using the off-the-shelf sensor alone, we were able to detect stressful events with an F1-score of up to 0.87 in the lab and 0.66 in the field, on par with clinical-grade sensors.

NeckSense: A Multi-Sensor Necklace for Detecting Eating Activities in Free-Living Conditions.

We present the design, implementation, and evaluation of a multi-sensor, low-power necklace, NeckSense, for automatically and unobtrusively capturing fine-grained information about an individual's eating activity and eating episodes, across an entire waking day in a naturalistic setting. NeckSense fuses and classifies the proximity of the necklace from the chin, the ambient light, the Lean Forward Angle, and the energy signals to determine chewing sequences, a building block of the eating activity. It then clusters the identified chewing sequences to determine eating episodes. We tested NeckSense on 11 participants with and 9 participants without obesity, across two studies, where we collected more than 470 hours of data in a naturalistic setting. Our results demonstrate that NeckSense enables reliable eating detection for individuals with diverse body mass index (BMI) profiles, across an entire waking day, even in free-living environments. Overall, our system achieves an F1-score of 81.6% in detecting eating episodes in an exploratory study. Moreover, our system can achieve an F1-score of 77.1% for episodes even in an all-day-long free-living setting. With more than 15.8 hours of battery life, NeckSense will allow researchers and dietitians to better understand natural chewing and eating behaviors. In the future, researchers and dietitians can use NeckSense to provide appropriate real-time interventions when an eating episode is detected or when problematic eating is identified.

Evaluating the Reproducibility of Physiological Stress Detection Models.

Recent advances in wearable sensor technologies have led to a variety of approaches for detecting physiological stress. Even with over a decade of research in the domain, there still exist many significant challenges, including a near-total lack of reproducibility across studies. Researchers often use some physiological sensors (custom-made or off-the-shelf), conduct a study to collect data, and build machine-learning models to detect stress. There is little effort to test the applicability of the model with similar physiological data collected from different devices, or the efficacy of the model on data collected from different studies, populations, or demographics. This paper takes the first step towards testing reproducibility and validity of methods and machine-learning models for stress detection. To this end, we analyzed data from 90 participants, from four independent controlled studies, using two different types of sensors, with different study protocols and research goals. We started by evaluating the performance of models built using data from one study and tested on data from other studies. Next, we evaluated new methods to improve the performance of stress-detection models and found that our methods led to a consistent increase in performance across all studies, irrespective of the device type, sensor type, or the type of stressor. Finally, we developed and evaluated a clustering approach to determine the stressed/not-stressed classification when applying models on data from different studies, and found that our approach performed better than selecting a threshold based on training data. This paper's thorough exploration of reproducibility in a controlled environment provides a critical foundation for deeper study of such methods, and is a prerequisite for tackling reproducibility in free-living conditions.