To perform a physical performance test is a good way to keep track of one’s health and can be beneficial to find evidence of deviations in the body. This thesis focuses on the development of a mobile-based heart rate algorithm that can be used with the Queens College Step Test, on the behalf of Mobistudy. Mobistudy wants to include such a test in their mobile application which aims to become a tool for researchers to use to gather data. The algorithm uses the mobile device’s camera to collect data from the user’s finger and uses that data to calculate the heart rate. The algorithm was first tested with data collected during the development and the results has an average error of less than 5% and a standard deviation of less than 3%. Two participants between the age of 20-25 performed three sets each of the Queens College Step Test and the results showed that the algorithm was accurate in its estimation of the heart rate after the test. / Genom att utföra ett test av ens fysiska prestanda kan man utvärdera ens hälsostatus och upptäcka indikationer på avvikelser i kroppen. Syftet med detta arbete är att utveckla en mobilbaserad algoritm som kan beräkna och uppskatta ens puls när man utför the Queens College Step Test på begäran av Mobistudy. Mobistudy vill inkludera detta test i deras mobilapplikation som fokuserar på att kunna användas som ett verktyg inom forskning för att samla in data. Algoritmen använder sig av mobilens kamera för att samla in data från användarens finger och använder den insamlade data för att beräkna pulsen. Algoritmen testades först gentemot data som samlades in vid utvecklingsstadiet och resultatet visade på att genomsnittliga felet var under 5% samt att standardavvikelsen var under 3%. Två deltagare mellan åldern 20 och 25 utförde tre tester var utav the Queens College Step Test och resultatet visade att algoritmen var tillräckligt noggrann i sin uppskattning av pulsen efter ett utfört test.

Master of Science / Department of Electrical and Computer Engineering / Steven Warren / Pulse oximetry, a noninvasive circulatory system monitoring technique, has been widely adopted in clinical and homecare applications for the determination of heart rate and blood oxygen saturation, where measurement locations are typically limited to fingertips and earlobes. Prior research indicates a variety of additional clinical parameters that can be derived from a photoplethysmogram (PPG), the fundamental time-domain signal yielded by a pulse oximeter sensor. The gap between this research potential and practical device applications can be decreased by improvements in device design (e.g., sensor performance and geometry, sampling fidelity and reliability, etc.) and PPG signal processing. This thesis documents research focused on a novel pulse oximeter design and the accompanying PPG signal processing and interpretation. The filter-free reflectance design adopted in the module supplements new methods for signal sampling, control, and processing, with a goal to acquire high-fidelity raw data that can provide additional physiologic data for state-of-health analyses. Effective approaches are also employed to improve signal stability and quality, including shift-resistant baseline control, an anti-aliasing sampling frequency, light emitting diode intensity autoregulation, signal saturation inhibition, etc. MATLAB interfaces provide data visualization and processing for multiple applications. A feature detection algorithm (decision-making rule set) is presented as the latest application, which brings the element of intelligence into the pulse oximeter design by enabling onboard signal quality verification. Two versions of the reflectance sensor were designed, built, calibrated, and utilized in data acquisition work. Raw data, which are composed of four channels of signals at a 240 Hz sampling rate and a 12-bit precision, successfully stream to a personal computer via a serial connection or wireless link. Due to the optimized large-area sensor and the intensity autoregulation mechanism, PPG signal acquisition from measurement sites other than fingertips and earlobes, e.g., the wrist, become viable and retain signal quality, e.g., signal-to-noise ratio. With appropriate thresholds, the feature detection algorithm can successfully indicate motion occurrence, signal saturation, and signal quality level. Overall, the experimental results from a variety of subjects and body locations in multiple applications demonstrate high quality PPGs, prototype reliability, and prospects for further research value.

No description available.

Blood pressure (BP) monitoring is an important tool for management of hypertension, which is a significant risk for cardiovascular disease and premature death. Since cuff-based BP measurement can be uncomfortable and does not provide continuous readings, several cuffless methods that are typically based on within-beat information or on the pulse transit time (PTT) have recently been investigated. This work proposes a novel cuffless BP estimation approach that mainly uses the information from cardiovascular dynamics of photoplethysmogram (PPG) waveforms. This work is divided into three parts. The first part proposes a calibration-free approach that uses dynamic changes in the pulse waveform. Results from 200 patients showed that the method achieved grade B, in terms of accuracy, for diastolic blood pressure (DBP) based on the British Hypertension Society (BHS) standard and complied with the accuracy requirements of the Association for Advancement of Medical Instrumentation/European Society of Hypertension/International Organization for Standardization (AAMI/ESH/ISO) standard. The second part presents a method based on calibrated cardiovascular dynamics, achieved through a mathematical model that relates reflective PTT (R-PTT) to BP. Results from 30 patients showed a mean error (ME) of 0.58 mmHg, standard deviation of the error (SDE) of 8.13 mmHg, and a mean absolute error (MAE) of 4.93 mmHg for DBP and an ME of 2.52 mmHg, SDE of 12.28 mmHg, and an MAE of 8.82 mmHg for systolic blood pressure (SBP). The third part proposes a calibration-free method that combines morphology features and dynamic changes of the pulse waveform over short intervals. In this method a neural network was trained on 200 patients and tested on never-seen data from 25 other patients and provided an ME of -0.31 mmHg, SDE of 4.89 mmHg, and MAE of 3.32 mmHg for DBP and an ME of -4.02 mmHg, SDE of 10.40 mmHg, and MAE of 7.41 mmHg for SBP. Overall, the results show that cardiovascular dynamics may contribute useful information for cuffless estimation of BP.

Exploiting photoplethysmography signals (PPG) for non-invasive blood pressure (BP) measurement is interesting for various reasons. First, PPG can easily be measured using fingerclip sensors. Second, camera based approaches allow to derive remote PPG (rPPG) signals similar to PPG and therefore provide the opportunity for non-invasive measurements of BP. Various methods relying on machine learning techniques have recently been published. Performances are often reported as the mean average error (MAE) on the data which is problematic. This work aims to analyze the PPG- and rPPG based BP prediction error with respect to the underlying data distribution. First, we train established neural network (NN) architectures and derive an appropriate parameterization of input segments drawn from continuous PPG signals. Second, we use this parameterization to train NNs with a larger PPG dataset and carry out a systematic evaluation of the predicted blood pressure. The analysis revealed a strong systematic increase of the prediction error towards less frequent BP values across NN architectures. Moreover, we tested different train/test set split configurations which underpin the importance of a careful subject-aware dataset assignment to prevent overly optimistic results. Third, we use transfer learning to train the NNs for rPPG based BP prediction. The resulting performances are similar to the PPG-only case. Finally, we apply different personalization techniques and retrain our NNs with subject-specific data for both the PPG-only and rPPG case. Whilst the particular technique is less important, personalization reduces the prediction errors significantly.

abstract: Advancements in mobile technologies have significantly enhanced the capabilities of mobile devices to serve as powerful platforms for sensing, processing, and visualization. Surges in the sensing technology and the abundance of data have enabled the use of these portable devices for real-time data analysis and decision-making in digital signal processing (DSP) applications. Most of the current efforts in DSP education focus on building tools to facilitate understanding of the mathematical principles. However, there is a disconnect between real-world data processing problems and the material presented in a DSP course. Sophisticated mobile interfaces and apps can potentially play a crucial role in providing a hands-on-experience with modern DSP applications to students. In this work, a new paradigm of DSP learning is explored by building an interactive easy-to-use health monitoring application for use in DSP courses. This is motivated by the increasing commercial interest in employing mobile phones for real-time health monitoring tasks. The idea is to exploit the computational abilities of the Android platform to build m-Health modules with sensor interfaces. In particular, appropriate sensing modalities have been identified, and a suite of software functionalities have been developed. Within the existing framework of the AJDSP app, a graphical programming environment, interfaces to on-board and external sensor hardware have also been developed to acquire and process physiological data. The set of sensor signals that can be monitored include electrocardiogram (ECG), photoplethysmogram (PPG), accelerometer signal, and galvanic skin response (GSR). The proposed m-Health modules can be used to estimate parameters such as heart rate, oxygen saturation, step count, and heart rate variability. A set of laboratory exercises have been designed to demonstrate the use of these modules in DSP courses. The app was evaluated through several workshops involving graduate and undergraduate students in signal processing majors at Arizona State University. The usefulness of the software modules in enhancing student understanding of signals, sensors and DSP systems were analyzed. Student opinions about the app and the proposed m-health modules evidenced the merits of integrating tools for mobile sensing and processing in a DSP curriculum, and familiarizing students with challenges in modern data-driven applications. / Dissertation/Thesis / M.S. Electrical Engineering 2013

Accurate assessment of a child’s health is critical for appropriate allocation of medical resources and timely delivery of healthcare in both primary care (GP consultations) and secondary care (ED consultations). Serious illnesses such as meningitis and pneumonia account for 20% of deaths in childhood and require early recognition and treatment in order to maximize the chances of survival of affected children. Due to time constraints, poorly defined normal ranges, difficulty in achieving accurate readings and the difficulties faced by clinicians in interpreting combinations of vital signs, vital signs are rarely measured in primary care and their utility is limited in emergency departments. This thesis aims to develop a monitoring and data fusion system, to be used in both primary care and emergency department settings during the initial assessment of children suspected of having a serious infection. The proposed system relies on the photoplethysmogram (PPG) which is routinely recorded in different clinical settings with a pulse oximeter using a small finger probe. The most difficult vital sign to measure accurately is respiratory rate which has been found to be predictive of serious infection. An automated method is developed to estimate the respiratory rate from the PPG waveform using both the amplitude modulation caused by changes in thoracic pressure during the respiratory cycle and the phenomenon of respiratory sinus arrhythmia, the heart rate variability associated with respiration. The performance of such automated methods deteriorates when monitoring children as a result of frequent motion artefact. A method is developed that automatically identifies high-quality PPG segments mitigating the effects of motion on the estimation of respiratory rate. In the final part of the thesis, the four vital signs (heart rate, temperature, oxygen saturation and respiratory rate) are combined using a probabilistic framework to provide a novelty score for ranking various diagnostic groups, and predicting the severity of infection in two independent data sets from two different clinical settings.

Photoplethysmogram (PPG) signals detect blood volume variations during the heart cycle. They are useful to track physiological parameters of an individual, such as heart rate, heart rate variability or oxygen saturation. They are typically obtained using smart wearables and pulse oximeters, but our goal is to create remote PPG (rPPG) signals from video cameras. Since the signals obtained from a video camera are the RGB channels, we carried out an empirical study of the performance of each channel. RGB channels can be used to generate rPPG signals, but also as input to other processes that do so. As reference ground truth, we use contact PPG (cPPG) readings from pulse oximeters in the fingertip. In terms of several metrics, including dynamic time warping (DTW), Pearson’s correlation coefficient, root mean squared error (RMSE), and Beats-per-minute Difference (|∆BPM|), the green channel produced the best results, followed by the blue and red channels. Despite the green channel consistently outperforming the blue and red channels, the outcomes varied greatly depending on the dataset. We also applied different methods to obtain rPPG signals from the RGB channels, including CHROM-based rPPG, local group invariance (LGI), and plane-orthogonal-to-skin (POS). These techniques were contrasted with our novel technique based on a machine learning approach. For that, we made use of a variety of architectures, including convolutional neural networks and long short-term memory. The results were favourable for the ML approach in terms of DTW, r and |∆BPM|. / Fotopletysmogram (PPG)-signaler upptäcker variationer av blodvolym under hjärtcykeln. De är användbara för att spåra fysiologiska parametrar för en individ, såsom hjärtfrekvens, hjärtfrekvensvariabilitet eller syremättnad. De erhålls vanligtvis med smarta bärbara sensorer och pulsoximetrar, men vårt mål är att skapa fjärr-PPG (rPPG)-signaler från videokameror. Eftersom signalerna erhållna från en videokamera är RGB -kanalerna genomförde vi en empirisk studie av prestandan för varje kanal. RGB-kanaler kan användas för att generera rPPG-signaler, men också som input till andra processer som gör det. Som referens använder vi kontakt-PPG (cPPG) avläsningar från pulsoximetrar i fingertoppen. När det gäller flera mätvärden, inklusive Dynamic Time Warping (DTW), Pearsons korrelationskoefficient, Root Mean Squared Error (RMSE) och Beats-Per-minut-skillnaden (|∆BPM|). Uppnåddes bästa resultat med den gröna kanalen, följt av de blå och röda kanalerna. Trots att den gröna kanalen konsekvent överträffade de blå och röda kanalerna varierade resultaten mycket beroende på datasetet. Vi använde också olika metoder för att erhålla rPPG-signaler från RGB-kanalerna, inklusive CHROM-baserad rPPG, lokal gruppinvarians (LGI) och plan-ortogonal-till-hud (POS). Dessa tekniker kontrasterades med vår nya teknik baserat på en maskininlärningsstrategi. För det använde vi en mängd olika arkitekturer, inklusive konvolutionella neurala nätverk och LSTM-nätverk. Resultaten var gynnsamma för ML-metoden när det gäller DTW, R och |∆BPM|.

No description available.

Blind source separation (BSS) aims at separating useful signal content from distortions. In the contactless acquisition of vital signs by means of the camera-based photoplethysmogram (cbPPG), BSS has evolved the most widely used approach to extract the cardiac pulse. Despite its frequent application, there is no consensus about the optimal usage of BSS and its general benefit. This contribution investigates the performance of BSS to enhance the cardiac pulse from cbPPGs in dependency to varying input data characteristics. The BSS input conditions are controlled by an automated spatial preselection routine of regions of interest. Input data of different characteristics (wavelength, dominant frequency, and signal quality) from 18 postoperative cardiovascular patients are processed with standard BSS techniques, namely principal component analysis (PCA) and independent component analysis (ICA). The effect of BSS is assessed by the spectral signal-tonoise ratio (SNR) of the cardiac pulse. The preselection of cbPPGs, appears beneficial providing higher SNR compared to standard cbPPGs. Both, PCA and ICA yielded better outcomes by using monochrome inputs (green wavelength) instead of inputs of different wavelengths. PCA outperforms ICA for more homogeneous input signals. Moreover, for high input SNR, the application of ICA using standard contrast is likely to decrease the SNR.