Artificial Intelligence (AI) enabledembedded devices are becoming increasingly important in the field of healthcarewhere such utilized to assist physicians, clinicians, and surgeonsin their diagnosis, therapy planning, rehabilitation. In current practice,designing models requires experts DNN design, healthcare, embeddedsystems. Additionally, task migrating a differentmicrocontroller (MCU) platform typically significant effort re-sizeand/or re-train model. This paper shows that Neural...

Smoking Cessation is a vital wellness application as smoking has health issues pertaining to cancer, cardio-pulmonary diseases, hypertension, and diabetes. This paper presents method for online on-device puff detection on microcontroller-based wearable device. We design specialized Convolutional Neural Network (CNN) based model from Respirational Inductance Photoplethysmogram (RIP) with 6-axis IMU signal achieving 81% F1-score provide an algorithm quantify episodes the certainty of detected...

Fruits provide essential nutrition in most natural form suitable for human beings. They are best when ripened naturally. However, industrialization has provided many ways quick ripening and extended shelf life of fruits. Detection artificial could be done by sophisticated methods like chemical analysis lab or visual inspection experts, which may not feasible all the time. Of fruits, banana is consumed fruit around world. Adulteration can have devastating effects on masses scale. It figured,...

In recent years, wearable devices and sensing systems have become an integral part of the deployment Artificial Intelligence (AI) on Edge. These intelligent edge perform first-level analytics to reduce data transfer, improve response latency, preserve privacy. Thus, there is a growing focus design machine learning models with high accuracy, low resource footprint, suitable for wide range applications in domains healthcare, wellness, lifestyle, among others. A majority these tasks comprise...

Stress management is paramount to maintaining optimal health and well-being; stress builds up in spikes, causing problems like hypertension anxiety, necessitating personalized interventions delivered real-time through wearable technology. This work underscores the pivotal role of unobtrusive detection presents development auto-generated compact models tailored for on-device inference Neural Architecture Search (NAS). These aim facilitate efficient monitoring directly on low-power devices,...

Smoking is a significant cause of death and deterioration health worldwide, affecting active passive smokers. Cessation smoking contributes to an essential wellness application owing the broad range problems such as cancer, hypertension, several cardiopulmonary diseases. Personalized smoking-cessation applications can be very effective in helping users stop if there are detections interventions done at right time. This requires real-time detection puffs. Such made feasible by day-long...

Wearable sensor-based stress detection is a well-explored area of research in the domain Affective Computing and can be performed with help non-invasive sensing modalities like Electrodermal Activity (EDA). The EDA sensors wearable form factor are commonly available on commercial off-the-shelf devices. In recent years, increased availability such devices to end-users, these applications have become more pervasive thus require greater level optimization for continuous usage...

We propose a practical Heart Rate Estimation algorithm utilizing wrist-based photoplethysmography (PPG) signals for continuous health monitoring of crane workers who spend long hours in an isolated cabin the harsh factory environment. Our novelty lies devising low footprint that can reliably estimate presence motion artefact as well offers feasibility deploying on wearable platform. More particularly, our solution addresses two fundamental issues: a) correcting weak wrist PPG signal from...

Cardiopulmonary disease prognosis can achieve therapeutic edge if the disorders be detected and attended to at an early stage. This work proposes 'Cardiopulmonary Care Platform (C2P)' which in its current initial phase, targets subjects stage of Functional Capacity II as per NYHA staging system by detecting physiological fatigue, signs dispnea palpitation, using a smartwatch after subject has undergone spell physically intensive activity. A novel computationally efficient solution is devised...

Continuous monitoring of cardiac health through single-lead wearable Electrocardiogram (ECG), is important for paroxysmal Atrial Fibrillation (AF) detection. Wearable ECG straps, watches, and implantable loop recorders (ILR) are based on this paradigm. These devices used by medical professionals to view data from multiple patients, perform continuous analysis provide immediate care patients. display simple screening alerts the subjects generate distress signals people working outdoors or in...

Customized, on-device applications that provide timely interventions about smoking episodes are very helpful for cessation. For this, real-time detection of puffs necessary through unobtrusive wearable devices. This work demonstrates auto-generated tiny puff models inference on low-power

Heart rate estimation from the wrist-based photo-plethysmography sensor possesses a great challenge due to its poor signal quality. The potential problems are categorized as different optical response of diverse skin tone, inconsistency in placement and off line calibration. In this paper, we propose system which attempts at personalizing wrist PPG calibration with joint objective maximizing quality minimizing energy consumption. Preliminary assessment demonstrates feasibility our proposed...

It is with great pleasure that we extend our warmest welcome to the inaugural edition of Workshop on AutoML in Pervasive Sensing Systems (AutoMLPerSys 2024).As chairs this workshop, are thrilled bring together experts and researchers, from around world explore latest advancements, challenges, opportunities automated machine learning (AutoML) within context pervasive sensing systems.We provide a platform for Academia Industry collaborate solutions computing systems, while exchanging insights...

Wrist based devices, like smart-watches, fitness bands and health monitors all provide a common sensor called Photoplethysmography (PPG) to measure optical pulse signal. This is usually used derive the instantaneous heart-rate (HR), which useful while doing any exercise or monitor on regular basis for chronic patients. However, one major issue with signal that it easily corrupted by ambulatory motion generated hand movements of subject. Since, these devices also come equipped an independent...

The growing need for artificial intelligence in multitude of domains has over the past few years triggered some concerns related to latency and reliability when huge amounts sensor data are moved from sensors cloud computation. These concerns, coupled with sustainability drives green computing, have emerged as key drivers move computation closer source, on devices which often less power-hungry compared machines at centers. However, close sources, or sensors, typically constrained terms...

Wearable health monitoring has become a very familiar term in today'sworld. One of the most popular means ofwearable sensing is photoplethysmogram (PPG). Due to its unobtrusive and ubiquitous nature, it gaining popularity among people everywhere. ease use, utility such technology increasing day by day. However, theworld researchers, accurate estimation heart rate (HR) presence motion artefacts remains an unsolved problem due susceptibility PPG signals corruption artefacts. The way which...

The paradigm of wellness consists both physical and mental wellness. One important parameter in the monitoring cardiac health during activity, which requires measurement ambulatory heart rate (HR). With advent smart wearable devices, heart-rate using Photoplethysmogram (PPG) has become a commodity. However, arriving at reliable real-time daily activities is an open research problem. In this paper, we propose method based on Weiner Filter, to estimate correct HR values presence motion, while...