Showing papers presented at "IEEE International Symposium on Medical Measurements and Applications in 2019"

Cardio Twin: A Digital Twin of the human heart running on the edge

Abstract: We present the Cardio Twin architecture for Ischemic Heart Disease (IHD) detection designed to run on the edge. We classify non-myocardial and myocardial conditions with a CCN. This CNN generates features from the electrocardiograms and performs the classification task. The database used is "PTB Diagnostic ECG Database" from Physio Bank and it comes from 200 different people. Each patient data sample was partitioned into 2.5 second windows for training. The implemented model achieved 85.77% accuracy and used 4.8 seconds for each sample classification. The results show that technology is ready to fully support demanding processes, such as Digital Twin, on the edge.

An Overview on Internet of Medical Things in Blood Pressure Monitoring

Abstract: The increasing pervasiveness of wearable sensors opens new scenarios in the continuous monitoring of health parameters. In particular, wearables are becoming the sensing part of the Internet of Medical Things (IoMT), i.e. IoT in the healthcare field. Currently, several IoMT based devices capable of measuring blood pressure (BP) are starting to be offered on the market, giving the possibility to monitor BP every time and everywhere. An open issue is the lack of the traceability and reliability of the BP measurements. The aim of this paper is stimulating the research to fulfil this lack by presenting an overview for IoMT and commercial wearables in BP monitoring, from a metrological point of view.

The architecture of an innovative smart T-shirt based on the Internet of Medical Things paradigm

Abstract: This paper deals with a description of an innovative Internet of Medical Things (IoMT) system for implementing personalized health services. The proposed IoMT system has the following advantages respect to the state-of-the-art systems available in literature: (i) it fuses the data provided by several sensors, inertial measurement unit, the bio-impedance and electrocardiogram, (ii) it uses Compressed Sensing (CS) of data prior to transmission, and (iii) it adopts distributed artificial intelligence at the edge for anomaly detection. A description of the specific features and requirements of the wearable device that will be embedded on a smart T-shirt is reported. According to the delineated requirements, an architecture for the wearable device is proposed. Finally the contribution of Artificial Intelligence in the proposed IoMT system is discussed, aiming at identifying anomalies and supporting the process of decision making in the early diagnosis of diseases both at individuals level (local knowledge) or groups of individuals level (global knowledge).

A Novel Method for Compressed Sensing based Sampling of ECG Signals in Medical-IoT era

Abstract: The paper presents a novel method for Compressed Sensing (CS) based sampling scheme, which is suitable for implementation in a real-time data acquisition system, for the Internet-of-Medical-Things (IoMT) era. The method relies on exploiting CS theory for bioelectrical signals, and in particular, in this paper, the electrocardiographic (ECG) type of signals are considered. The novelty of the method consists of designing the sensing matrix considering the information that is related to the auto-correlation coefficients from the observed signal. Thus, the mathematical model for CS based acquisition of ECG signals and its feasibility for practical implementation for an IoMT based system is described. The results presented herein clearly demonstrates the robustness and usefulness of the proposed method.

Towards Privacy-Preserving Deep Learning based Medical Imaging Applications

Abstract: Following the reports of breakthrough performances, machine learning based applications have become very popular in the medical field. However, with the recent increase in concerns related to data privacy, and the publication of specific regulations (e.g. GDPR), the development and, thus, exploitation of deep learning based applications in clinical decision making processes, has been rendered impossible in many cases. Herein, we describe and evaluate an approach that employs Fully Homo-morphic Encryption for allowing computations to be performed on sensitive data. Specifically, the solution exploits the MORE scheme and does not disclose patient data. The chosen encryption scheme increases the runtime only marginally and, importantly, allows for operations to be performed directly on floating point numbers, which represents a critical property for artificial neural networks. The feasibility and performance are first evaluated on a standard benchmarking application (MNIST digit classification). Next, we considered a medical imaging application, i.e. classification of coronary views in X-ray angiography. The reported results indicate that the proposed solution has great potential: (i) computational results are indistinguishable from those obtained with the unencrypted variants of the deep learning based applications, and (ii) run times increase only marginally. Finally, we also discuss in detail security concerns, and emphasize that the proposed solution may be employed in several practical applications, while still significant limitations remain to be solved in future work.

ECG WATCH: a real time wireless wearable ECG

Abstract: Cardiovascular diseases (CVD) remain the most common cause of death worldwide. Standard techniques such as 12-leads electrocardiogram (ECG) or Holter system are not sufficient to fully address some sporadic ECG anomalies like atrial fibrillation. Several low-cost wearable devices have already been proposed but, each of them misses some important features. The ECG WATCH has been developed to target these problems. It is a non-expensive, wearable, easy-to-use health device to monitor CVD patients heart activity anytime, anywhere, without the need to go physically to hospitals or cardiologists. Recordings need just 10 seconds; it also embeds an algorithm to detect possible atrial fibrillation episodes.

Investigation of Wearable Motion Capture System Towards Biomechanical Modelling

Abstract: Motion Capture (MOCAP) systems provide kinematic data as an output. Biomechanical models use this data as an input to calculate the forces acting on the joints of human body. For high bio-fidelity modelling, accurate daily life data are required. Here, we asses a novel marker-less wearable MOCAP system, SmartSuit Pro (Rokoko, Copenhagen, Denmark) for feeding in data to a biomechanical modelling software. This suit is utilised for medical data collection purposes for the first time. The study provides proof of concept for the upper body motion in vitro using 5th percentile female skeleton model. Flexion -extension movements of the shoulder and elbow were simulated using controlled step motor to quantify the deviation between the planned and the measured profile by the wearable suit. Cross validation is completed in vivo using Optitract (LEYARD,Corvallis, USA). In vivo data was collected from healthy volunteers with no previous history of upper extremity disorder. For repeatability and reliability purposes, 3 sets of motions were repeated 3 times to measure flexion/extension range of motion (RoM). The relative peak angles were calculated. Root Mean Square Error (RMSE) was as 0.46° and 0.31° respectively for shoulder and elbow. In vivo, RMSE values for shoulder and elbow flexion were 0.66° and 0.51° respectively. Pearson correlation coefficient was calculated as 0.7° and 0.77° for shoulder and elbow between OptiTrack and Smartsuit Pro respectively. Bland - Altman plots showed that Rokoko system produces data comparable to OptiTrack. The collected data was fed into Biomechanics of Bodies (BoB) simulation software for calculating kinetic data. In conclusion, Rokoko system could be an alternative solution while measuring RoM for biomedical purposes.

TL-FALL: Contactless Indoor Fall Detection Using Transfer Learning from a Pretrained Model

Abstract: A new radar-based fall detection method is proposed using the recent advances in deep neural networks. An ultrawideband radar is used to monitor human daily activities and identify the occurrence of falls. A transfer learning approach is employed based on a pre-trained model on ImageNet dataset to realize a robust feature extraction from radar data. The architecture and depth of the model are fine-tuned to radar time-frequency representations. From the results, it is observed that the proposed transfer learning based method can achieve a detection accuracy for fall incidents higher that those of the other methods.

Cardiac monitoring with a smart textile based on polymer-encapsulated FBG: influence of sensor positioning

Abstract: In recent years, wearables are exploding in popularity as unobtrusive devices able to extend traditional healthcare delivery systems. Smart textiles are one of the main innovative types of wearables used for non-invasive and continuous monitoring of cardiac activity. A prominent solution is based on the detection of vibrations induced on the chest surface by the heart beating (i.e., precordial motions). In the literature, different sensor positions have been investigated, but it appears to be a lack of accepted standard points for the detection of heart-induced motions. In this work, a smart textile based on fiber Bragg grating (FBG) sensor has been proposed to detect the precordial motions on the chest. The feasibility of the smart textile for cardiac monitoring has been evaluated on three volunteers at three measurement points. Then, the influence of the measurement site on the response of the smart textile has been preliminarily assessed in terms of peak-to-peak amplitude of the signal. The signal amplitude is greater than the noise, so it allows detecting precordial motions. These promising results foster future investigations on the capability and performance of the system in estimating heart rate. Further tests will also be devoted to finding out the optimal measurement points to standardize the sensors positioning in this specific application.

In-Hand Telemanipulation Using a Robotic Hand and Biology-Inspired Haptic Sensing

Abstract: A new approach to the in-hand manipulation problem, using fuzzy controlled teleoperation and a recently proposed multimodal tactile sensor are presented in this article. Real-time grasping experiments were performed by a fuzzy controlled gripper, with a compliant tactile sensor mounted on the gripper. As a new approach for in-hand manipulation and robot dexterity, a teleoperated thumb performed object rotation while the gripper maintained stable grasping. The data from the experiments were consistent and, thus, provided a substantial tool for in-hand control algorithms.

Nb 2 O 5 thin film-based conductometric sensor for acetone monitoring

Abstract: Nowadays, effective detection of gases at ppm and ppb level is of crucial importance in a wide range of applications, such as industrial processes, environmental monitoring, public security and medical investigation. Several sensor types have been developed in last decades, among them the metal oxide gas sensors are the most promising for low-cost and portable applications, where good sensitivity and selectivity, together with small size are important constraints. The proposed conductometric gas sensor has been manufactured depositing a Nb 2 O 5 thin film by plasma sputtering on a commercial alumina substrate with platinum interdigitated electrodes and heater which size is 3 mm × 6 mm with a thickness of 1 mm. The Nb 2 O 5 thin film has been characterized by FE-SEM and XPS analysis. The sensor performance towards several target gases have been evaluated employing an experimental setup specifically developed for the characterization of gas sensors. The sensor exhibits good sensitivity towards acetone, a bio-marker found in human breath of diabetes patients. This makes the sensor promising in the noninvasive diagnosis of this kind of disease.

MCNP-FBSM: Development of MCNP/MCNPX Source Model for Simulation of Multi-Slice Fan-Beam X-Ray CT Scanners

Abstract: Computed tomography (CT) is one of the most valuable diagnostic imaging tools in the clinic and is widely used worldwide. One of the main motivations driving research and development in CT is to achieve better image quality while keeping the radiation dose to the patient as low as possible. In this regard, computer simulations play a key role in the optimization of CT design. In this work, a fan-beam source model (FBSM) for the simulation of multi-slice fan-beam CT scanners using the MCNP Monte Carlo code, has been developed and implemented. The use of this model removes the need for using the collimator in the system configuration and thus to overcome the perennial problem of particle starvation imposed by the collimator. The accuracy of our developed MCNP-FBSM model was evaluated through comparison with previously published experimental results demonstrating good agreement. Therefore, the MCNP-FBSM based CT simulator is a powerful tool for protocol design, optimization of geometrical design parameters, assessment of image reconstruction algorithms and evaluating future innovations to improve the performance of CT scanners.

WiFi Channel State Information-Based Recognition of Sitting-Down and Standing-Up Activities

Abstract: Real-time recognition of human activities is an important functionality of smart spaces. It allows a wide range of security and healthcare applications. In this work, we use the Channel State Information (CSI) of WiFi signals to assess the patterns associated with dynamic human activities, including sitting-down and standing-up actions. We preprocess raw signals with both a Hampel filter and low-pass filter. The signals are then segmented into 20-packet labelled sequences. Features including kurtosis, maximum, mean, minimum, maximum peak, skew, standard deviation, and variance are extracted for each sequence, providing feature vectors of 168 variables to enable activity recognition. Features are normalized and a series of classifiers were trained and compared to predict three activity classes: stationary (seated or standing still), sitting-down, and standing-up. Preliminary results on data collected for a single subject achieve a classification accuracy of 98.4% with a medium Gaussian Support Vector Machine (SVM) to distinguish between these three classes.

Performance analysis of an electromagnetic tracking system for surgical navigation

Abstract: This paper describes the analysis and the improvement of an electromagnetic tracking system (EMTS) for surgical survey. The system is based on inductive coupling between five transmitting coils and a small coil censor, in order to overcome line-of-sight dependence.The performance of the control loop of coils excitation currents is investigated, ruling out coupling side effects which may potentially affect system’s performance.The results show a significant reduction of both voltage and current error, which may result in a more accurate position estimation.

Methodological and Measurement Concerns of Local Pulse Wave Velocity Assessment

Abstract: Local pulse wave velocity (local PWV) provides localized information on the stiffness of a particular segment of the heterogeneous arterial bed. As a screening tool for a host of cardiovascular events and hypertension, high accuracies may not be required for the local PWV measures. However, such high accuracies are required when the yielded local PWV measures are directly used for calculations of other biomechanical and physiological parameters. There are several techniques for measuring local PWV and each possesses its own methodological challenges and considerations for ensuring accuracy. This work focuses on the salient methodological and measurement concerns of the transit time-based local PWV evaluation techniques while pinpointing the ones that are equally applicable to the other approaches. To demonstrate these concerns, we have used our extensively validated image-free ultrasound technology – ARTSENS®, for the measurement of local PWV. We have conducted an in-vivo study on 20 subjects to investigate and comment on various hardware and software aspects that affect the measurement accuracy. Firstly, non-invasive sensing modalities that capture blood pulse waveforms uncorrupted by the influence of intervening tissue layers may be chosen. Inter-channel delay should be carefully quantified and eliminated as the transit time measures for an arterial segment smaller than 50 mm, are of the orders less than 20 ms. The presented study results demonstrate that the blood pulse waveforms should be sampled at high rates, to ensure adequate temporal resolution for performing transit time measurements of such small orders. Further, it was observed that the specifications of the bandpass filters used to process the blood pulse signals effect the accuracy of the measurements. The higher cut-off frequency of the band pass filter used is suggested to be in the range of 10 – 16 Hz and the filter order to be greater than 2 to achieve an RMSE smaller than 0.5 m/s.

Using Inertial Sensors to Evaluate Exercise Correctness in Electromyography-based Home Rehabilitation Systems

Abstract: Home-based rehabilitation systems can speed up recovery by enabling patients to exercise at home between rehabilitation sessions. However, home-based rehabilitation systems need to monitor and feedback exercises appropriately, as incorrect or imperfect exercises negatively impact the recovery of the patient. This paper describes a methodology for assessing the quality of rehabilitation exercises using inertial sensors, for a system that tracks exercises using surface electromyography sensors. This duality extends the information provided by the electromyography system since it provides a more comprehensive evaluation of posture and movement correctness. The methodology was evaluated with 17 physiotherapy patients, obtaining an average accuracy of 96% in detecting issues in the exercises monitored. The insights of this work are a first step to complement an electromyography-based home system to detect issues in movement and inform patients in real time about the correctness of their exercises.

Enanced video heart rate and respiratory rate evaluation: standard multiparameter monitor vs clinical confrontation in newborn patients

Abstract: Vital signs in neonatal care are usually measured by a multiparameter monitor (MM). However, this kind of device is inaccurate on the assessment in newborns, which can present different respiratory patterns and superficial breathing, resulting hardly detectable by the MM. In some cases, the assessment of the physician is more effective on the evaluation of the patient. In this paper we present a non-contact solution to measure the respiration rate (RR) and heart rate (HR) using a commercially available web-camera (WeC) and a personal computer (PC). We use Eulerian Video Magnification (EVM) to amplify the color variations in the image sequence. By extracting the signal from the thorax portion of the patients we are able to measure the RR and HR by spectral analysis. The results are compared with the MM and the assessment of the physician. The measures of RR and HR correlate with the data from the MM and result even more accurate than the MM when compared with the physician’s evaluation. We collect data on 40 patients demonstrating the feasibility of this method. The measure of RR and HR shows a root mean square error of 6.8 bpm for the HR and 2.1 bpm for the RR.

Comparison of two methods for estimating respiratory waveforms from videos without contact

Abstract: Among all the vital signs and physiological parameters, the respiratory rate (f R ) is still considered the neglected vital sign. In the clinical scenario, occupational scenario and during sport and activities different methods can be used to record the respiratory waveform and thus the f R .In this paper, we investigated the performances of two different methods for the non-contact monitoring of respiratory waveform by using RGB video signal acquired from a single built-in high-definition webcam. Two different methods have been tested: the first based on the intensity change of video pixels and the second on the extraction of the optical flow from the video sequence.The results obtained so far from eight subjects in the time and frequency domain reveal that optical flow method is better performing than its counterpart. The maximum percentage error for the optical flow method in the frequency domain to calculate the breathing rate stands at less than 3% for all the subjects investigated by far. Likewise, in the time domain, the maximum mean absolute error (MAE) reported for the optical flow method stands at less than 1% in all the subjects.

Accuracy of Heart Rate Measurements by a Smartwatch in Low Intensity Activities

Abstract: Wearable technologies to measure heart rate have improved remarkably in the last few years, making it possible for the users to find low-cost commercially available devices for easy and unobtrusive daily use. Despite the technological advancements, however, the accuracy provided by these devices in measuring the heart rate remains questionable, especially in scenarios not involving fitness or physical exercise monitoring. This paper concerns the metrological characterization of a Samsung Gear FIT SM-R350 smartwatch in measuring the user’s heart rate, focusing on the accuracy provided in low-intensity activities, by designing a suitable measurement protocol. The results obtained show an acceptable correlation of the smartwatch measurements with a certified pulse-oximeter, which suggest future activities aimed at testing the device’s accuracy in uncontrolled operative conditions and indoor/outdoor environments.

Measuring changes in gait kinematics due to walking-related fatigue in patients with Multiple Sclerosis

Abstract: The objective quantification of walking-related fatigue in patients with Multiple Sclerosis is still an open challenge. This study aims at measuring changes in gait kinematics due to fatigue in MS patients performing a 6-minute walking test (6MWT) and comparing the results with a control group. Nine MS patients and twenty-six healthy subjects were enrolled in the experimental protocol, consisting in the execution of a 6MWT along a 15-m pathway at self-selected speed. Data of lower limb segments were gathered by seven inertial sensors placed on pelvis, thighs, shanks and feet. The evolution of the range of motion (ROM) related to hip, knee and ankle joint was evaluated by partitioning the data related to the 6MWT into six time subgroups of 60 s. Then, maximum difference for each ROM was computed as the difference between the maximum and minimum related to each stride within all six subgroups and then averaged across strides; in addition, the ROM decrements between the first minute and the remaining five were also calculated. Statistical tests were performed to evaluate differences between MS cohort and control group. Results revealed significant effects of walking-related fatigue on gait kinematics of MS patients, as suggested by the continuous slowing down of the ROM during the execution of the experimental task for all joints. The here proposed synthetic indices can be considered as useful tools for the objective assessment of fatigue in MS patients in order to design specific treatments to reduce fatigability phenomenon.

Accuracy Enhancement of Total Force by Capacitive Insoles

Abstract: A coherent understanding of plantar pressures is crucial in the field of sports research and clinical studies. Real-time and dynamic analysis of plantar pressures require sensors that are in a wearable form factor. Sensor insoles built on a variety of principles satisfy this condition. The accuracy of total force provided by commercially available sensor insoles is significantly low compared to that provided by force plates, the gold standard for force measurements. The current work discusses the limitations of insole based force measurement from the design standpoint alone. We further propose a learning-based model to improve the accuracy of the capacitive sensor based insoles for the task of total force measurement. A learning-based model is developed by training on data collected from sensor insoles for loads varying from 0 to 80 kg on each insole. The absolute mean percentage error of left and right soles of specific sized capacitive insoles as compared to the gold standard has reduced from 8.01 to 5.19 and 35.85 to 21.64 respectively. The learning based model has shown promising results providing a reliable scope towards the extension of the work under variable foot anatomies and pressure distributions.

Prevalent Approach of Learning Based Cuffless Blood Pressure Measurement System for Continuous Health-care Monitoring

Abstract: For monitoring and controlling hypertension continuous measurement of blood pressure is very much needed, which can be feasible by recent technological advances of wearable devices which replaces traditional methods of blood pressure measurement. Continuous monitoring can provide precious data of individual health conditions. This work focuses on comparative study of previous methods with learning based approach for continuous blood pressure measurement system based on only photoplethysmograph (PPG). We have described the conventional methods of blood pressure measurement with their limitations; learning based feature exploration methods for blood pressure. And we concluded with result and few suggestions as a future work for estimating cuffless BP continuously. This work proposes a novel learning approach with 11 features estimated from 50 patients data collected from MIMIC-II database. The proposed model perceive stable and accurate measurement of blood pressure. Results showed that mean absolute error for systolic blood pressure (SBP) and diastolic blood pressure (DBP) are 12.62, 11.86, 12.70 and 3.78, 3.36, 3.57 in linear regression, support vector machine (SVM) and Gaussian regression respectively and it also shows 6.17, 6.21, 6.08 for mean arterial pressure (MAP).

Acoustical Characterization of Tissue - Mimicking Materials

Abstract: Ultrasonic devices are today widely used in hospitals for diagnosis and treatment purposes. Quality controls of ultrasonic devices used in the field of health must be performed on phantoms that simulate human tissue with similar acoustic properties to human tissue in terms of both safety and proximity to reality. The phantoms are very important materials and widely used in the calibration of ultrasound devices, performance tests, improvement of signal and noise ratios of existing systems, practical application of the users before the device use and interpretation of the images taken during the application and it is used in the characterization of medical ultrasonic systems since the 1960s. In addition, these phantoms are the reference material for calibration operations. In this study, acoustic characterization measurements were made for three different tissue-mimicking materials and the results obtained were compared with each other.

Apnea Event Detection Methodology using Pressure Sensors

Abstract: A method is proposed in this paper for the detection of suspected central apnea events from nocturnal data measured with pressure sensor arrays. Optimized set of time and frequency measures computed from overlapping segments of 9 s are fed to a support vector machine-based classifier to identify the possible origin of the segments, i.e., not-apneic or apneic episodes. The classifier decision on the sequence of successive segments is then used to detect a complete event. The classifier accuracy for the test data-set and the overall F-score of the system is found to be 94.43% and 74.44%, respectively.

Deep Network for Capacitive ECG Denoising

Abstract: Continuous monitoring of cardiac health under free living condition is crucial to provide effective care for patients undergoing post operative recovery and individuals with high cardiac risk like the elderly. Capacitive Electrocardiogram (cECG) is one such technology which allows comfortable and long term monitoring through its ability to measure biopotential in conditions without having skin contact. cECG monitoring can be done using many household objects like chairs, beds and even car seats allowing for seamless monitoring of individuals. This method is unfortunately highly susceptible to motion artifacts which greatly limits its usage in clinical practice. The current use of cECG systems has been limited to performing rhythmic analysis. In this paper we propose a novel end-to-end deep learning architecture to perform the task of denoising capacitive ECG. The proposed network is trained using motion corrupted three channel cECG and a reference LEAD I ECG collected on individuals while driving a car. Further, we also propose a novel joint loss function to apply loss on both signal and frequency domain. We conduct extensive rhythmic analysis on the model predictions and the ground truth. We further evaluate the signal denoising using Mean Square Error(MSE) and Cross Correlation between model predictions and ground truth. We report MSE of 0.167 and Cross Correlation of 0.476. The reported results highlight the feasibility of performing morphological analysis using the filtered cECG. The proposed approach can allow for continuous and comprehensive monitoring of the individuals in free living conditions.

Dielectric Measurements of Brain Tissues with Alzheimer’s Disease Pathology in the Microwave Region

Abstract: Relative permittivity and conductivity of brain tissue samples with severe form of Alzheimer’s disease was measured and compared to those of healthy brain tissues. The brain tissue samples contained a substantial amount of beta-amyloid plaques and tau tangles, both of which are key indicators of Alzheimer’s disease in a patient. The experiments were performed at frequencies from 20 MHz to 3 GHz when samples were defrosted at 5°C using a vector network analyzer and dielectric probes. A software was used to capture the dielectric properties and analyzed further. Two main categories of tissue samples were considered: the grey matter and white matter of the brain. Both of these categories were from the frontal cortex of the brain. In both categories, certain distinctions have been found in the measured dielectric properties that differ from healthy human brain tissues. The results from this novel study indicates a key characteristic that can be exploited in detecting plaques and tangles in the brain. In addition, this result indicates that radio frequency imaging techniques can potentially be used as a diagnostic method for detecting and monitoring Alzheimer’s disease.

Long-term, Ambulatory Respiratory Monitoring of COPD Patients using Garment-Adhered Sensors

Abstract: Management of chronic pulmonary diseases currently lacks longitudinal, high-compliant respiratory datasets. Such datasets have the potential to provide insight into the behavioral and physiological factors that predict disease progression and deterioration. Producing such datasets requires devices that can withstand real-world use and yet are unobtrusive enough to yield high patient adherence. This article first describes a set of necessary characteristics for a device to address these issues, informed by stakeholders in pulmonary health monitoring. It then describes Health Tags, a cardiorespiratory sensing platform whereby each device in a set is adhered to a patient’s undergarments to ease patient adherence. The paper reports on the longitudinal use of Health Tags among high-risk COPD patients in the home environment. There, Health Tags were worn an average of 78.7% of each 24-hour cycle and 87.1% of daytime hours during an average enrollment period of 35.5 days. No sign of a novelty effect was observed. The results are encouraging for the creation and application of clinically accurate, longitudinal, continuous pulmonary datasets. Such datasets can assist clinicians to improve care and help researchers generate and explore new endpoints related to respiratory disease.

A Novel Method for Measuring the MTF of CT Scanners: A Phantom Study

Abstract: The modulation transfer function (MTF) is well known as a crucial parameter in quality assurance of computed tomography (CT) scanners, which provides detailed information of both contrast and resolution of CT images. Different methods have been introduced and developed to calculate the MTF of CT scanners. However, a robust methodology which accurately estimates the MTF of CT scanners under the use of every range of object electron density and tube current-time product (mAs) has not been reported so far. To this aim, a new wavelet-based circular edge method for MTF measurement has been presented in this work. Owning to the edge spread function (ESF) susceptibility to noise, the approach was based on the assumption that the ESF can be decomposed into approximate and detailed information containing different noise levels. To evaluate the performance of our method, an in-house fabricated phantom containing various disk objects covering a range of electron densities from low to high values was scanned by a volumetric 64-slice clinical CT scanner under a range of low tube current from 50 to 100 mAs where image noise levels are higher than those of normal-dose scan protocols. Measurements have shown that our proposed method yielded an accurate estimation of MTF for high-density as well as low-density disk objects and it is valid and stable over the variations of noise levels.

Streaming Pressure Sensor Data for Cloud Processing

Abstract: The smart home that contains ambient and worn sensors to allow the assessment of older adult well-being has been the focus of many research projects and has the promise to enable increased independence for older adults and delay their need for institution-based care. Many of these projects have focused on the sensor technologies and have been performed in small scale pilots or tests with local data capture and storage. For wide scale deployment to large populations, a solution is required to allow these systems to be deployed that is low cost while also providing effective streaming of the data into the cloud for processing. In this paper, the performance of a Raspberry PI embedded computer being used with pressure sensitive mats and direct streaming to the IBM Cloud IoT service is presented. The paper shows this single sensor application consumes up to 30% of a single CPU core capacity of the platform for the task of flowing the data. The data also shows the variance in the capacity demands frequently drives core utilization to between 60 and 100%. The result is that a single core must be dedicated to the real-time data flow task as the data flow application (Node-Red) must run within a single core. The result indicates that the low cost Raspberry PI platform can meet the performance needs and the CPU capacity of other cores could allow additional processing of the data prior to transmission.

Ballistocardiographic Coupling of Triboelectric Charges into Capacitive ECG

Abstract: Capacitive ECG (cECG), a technique older than 50, is able to replace the gold standard ECG only in certain applications where unobtrusiveness and conformity are aimed at the expense of reduced signal quality. Triboelectric surface charges, motion artifacts, and resulting time-variant coupling capacitances are among the reasons for the signal deformations in cECG. In this paper, the mechanical vibrations of the human body are proposed as the source of time-variant coupling capacitances causing motion artifacts, which questions the applicability of adaptive filtering approaches to the problem of time-variant coupling capacitance. Ballistocardiogram (BCG) measurements on cECG electrodes are recorded and analyzed to investigate how these mechanical vibrations reflect on a differential biopotential measurement. Furthermore, using measured signals in a human experiment, numerical and test bench simulations were conducted to replicate how triboelectric surface charges might cause deformation on cECG signals.