Non-Contact Measurement of Blood Oxygen Saturation Using Facial Video Without Reference Values.

The continuous measurement of percutaneous oxygen saturation (SpO2) enables diseases that cause hypoxemia to be detected early and patients' conditions to be monitored. Currently, SpO2 is mainly measured using a pulse oximeter, which, owing to its simplicity, can be used in clinical settings and at home. However, the pulse oximeter requires a sensor to be in contact with the skin; therefore, prolonged use of the pulse oximeter for neonates or patients with sensitive skin may cause local inflammation or stress due to restricted movement. In addition, owing to COVID-19, there has been a growing demand for the contactless measurement of SpO2. Several studies on measuring SpO2 without contact used skin video images have been conducted. However, in these studies, the SpO2 values were estimated using a linear regression model or a look-up table that required reference values obtained using a contact-type pulse oximeter. In this study, we propose a new technique for the contactless measurement of SpO2 that does not require reference values. Specifically, we used certain approaches that reduced the influence of non-pulsating components and utilized different light wavelengths of video images that penetrated subcutaneously to different depths. We experimentally investigated the accuracy of SpO2 measurements using the proposed methods. The results indicate that the proposed methods were more accurate than the conventional method.

Prediction and prevention system for Severe Acute Respiratory Syndrome CoronaVirus 2 infection by preempting the onset of a cough.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is fast becoming one of the most significant infections worldwide. Of all the causes of SARS-CoV-2 infection, airborne-droplet infection via coughing is the most common. Therefore, if predicting the onset of a cough and preventing infection were possible, it would have a globally positive impact. Here, we describe a new prediction and prevention system for SARS-CoV-2 infection. Usually, air is inhaled prior to coughing, and the cough, which contains droplets of the virus, then occurs during acute exhalation. Therefore, if we can predict the onset of a cough, we can prevent the spread of SARS-CoV-2. At Tohoku University, a diagnosis system for evaluating swallowing motions and peripheral circulation has already been developed, and our prediction system can be integrated into this system. Using three-dimensional human body imaging, we developed a prediction system for preempting the onset of a cough. If we can predict the onset a cough, we can prevent the spread of SARS-CoV-2 infection, by decreasing the shower of virally active airborne droplets. Here, we describe the newly developed prediction and prevention system for SARS-CoV-2 infection that preempts the onset of a cough.Clinical Relevance- If predicting the onset of a cough and preventing infection were possible, it would have a globally positive impact. Here, we describe the newly developed prediction and prevention system for SARS-CoV-2 infection.

Comprehensive evaluation of machine learning algorithms for predicting sleep-wake conditions and differentiating between the wake conditions before and after sleep during pregnancy based on heart rate variability.

Introduction: Perinatal women tend to have difficulties with sleep along with autonomic characteristics. This study aimed to identify a machine learning algorithm capable of achieving high accuracy in predicting sleep-wake conditions and differentiating between the wake conditions before and after sleep during pregnancy based on heart rate variability (HRV). Methods: Nine HRV indicators (features) and sleep-wake conditions of 154 pregnant women were measured for 1 week, from the 23rd to the 32nd weeks of pregnancy. Ten machine learning and three deep learning methods were applied to predict three types of sleep-wake conditions (wake, shallow sleep, and deep sleep). In addition, the prediction of four conditions, in which the wake conditions before and after sleep were differentiated-shallow sleep, deep sleep, and the two types of wake conditions-was also tested. Results and Discussion: In the test for predicting three types of sleep-wake conditions, most of the algorithms, except for Naïve Bayes, showed higher areas under the curve (AUCs; 0.82-0.88) and accuracy (0.78-0.81). The test using four types of sleep-wake conditions with differentiation between the wake conditions before and after sleep also resulted in successful prediction by the gated recurrent unit with the highest AUC (0.86) and accuracy (0.79). Among the nine features, seven made major contributions to predicting sleep-wake conditions. Among the seven features, "the number of interval differences of successive RR intervals greater than 50 ms (NN50)" and "the proportion dividing NN50 by the total number of RR intervals (pNN50)" were useful to predict sleep-wake conditions unique to pregnancy. These findings suggest alterations in the vagal tone system specific to pregnancy.

Heart Rate Information-Based Machine Learning Prediction of Emotions Among Pregnant Women.

In this study, the extent to which different emotions of pregnant women can be predicted based on heart rate-relevant information as indicators of autonomic nervous system functioning was explored using various machine learning algorithms. Nine heart rate-relevant autonomic system indicators, including the coefficient of variation R-R interval (CVRR), standard deviation of all NN intervals (SDNN), and square root of the mean squared differences of successive NN intervals (RMSSD), were measured using a heart rate monitor (MyBeat) and four different emotions including "happy," as a positive emotion and "anxiety," "sad," "frustrated," as negative emotions were self-recorded on a smartphone application, during 1 week starting from 23rd to 32nd weeks of pregnancy from 85 pregnant women. The k-nearest neighbor (k-NN), support vector machine (SVM), logistic regression (LR), random forest (RF), naïve bayes (NB), decision tree (DT), gradient boosting trees (GBT), stochastic gradient descent (SGD), extreme gradient boosting (XGBoost), and artificial neural network (ANN) machine learning methods were applied to predict the four different emotions based on the heart rate-relevant information. To predict four different emotions, RF also showed a modest area under the receiver operating characteristic curve (AUC-ROC) of 0.70. CVRR, RMSSD, SDNN, high frequency (HF), and low frequency (LF) mostly contributed to the predictions. GBT displayed the second highest AUC (0.69). Comprehensive analyses revealed the benefits of the prediction accuracy of the RF and GBT methods and were beneficial to establish models to predict emotions based on autonomic nervous system indicators. The results implicated SDNN, RMSSD, CVRR, LF, and HF as important parameters for the predictions.

A Deep Learning Aided Drowning Diagnosis for Forensic Investigations using Post-Mortem Lung CT Images.

Feasibility of computer-aided diagnosis (CAD) systems has been demonstrated in the field of medical image diagnosis. Especially, deep learning based CAD systems showed high performance thanks to its capability of image recognition. However, there is no CAD system developed for post-mortem imaging diagnosis and thus it is still unclear if the CAD system is effective for this purpose. Particulally, the drowning diagnosis is one of the most difficult tasks in the field of forensic medicine because findings of the post-mortem image diagnosis are not specific. To address this issue, we develop a CAD system consisting of a deep convolution neural network (DCNN) to classify post-mortem lung computed tomography (CT) images into two categories of drowning and non-drowning cases. The DCNN was trained by means of transfer learning and performance evaluation was conducted by 10-fold cross validation using 140 drowning cases and 140 non-drowning cases of the CT images. The area under the receiver operating characteristic curve (AUC-ROC) for the DCNN was achieved 0.88 in average. This high performance clearly demonstrated that the proposed DCNN based CAD system has a potential for post-mortem image diagnosis of drowning.

Comparison of Visible and Infrared Video Plethysmography Captured from Different Regions of the Human Face.

Recently, video plethysmography (VPG) - a heart rate estimation technique using a video camera - has gained significant attention. Most studies of VPG have used a visible RGB camera; only a limited number of studies investigating near-infrared light (wavelength 750-2500 nm), which can be used even in a dark environment, have been performed. The purpose of this study was to investigate the differences between VPG data collected using visible light (VPGVIS) or near-infrared light (VPGNIR) from four facial areas (forehead, right cheek, left cheek, and nose). An experiment was conducted to obtain both VPGVIS and VPGNIR simultaneously by alternately irradiating the face with NIR and VIS lights. Experimental results showed that the root mean squared error of heart rate estimated using VPGNIR was 1 bpm higher than that of VPGVIS. However, contrary to our expectations, the power of the heartbeat-related component included in VPGNIR was not reduced despite the absorbance of hemoglobin in the NIR light range being 1/100 of that in the VIS light range. This result supports the hypothesis that a main factor in the generation of VPG waves was change in the optical properties caused by blood vessels compressing the subcutaneous tissue and the venous bed. Additionally, the accuracy of the heart rate estimation using VPG tended to be high when the nose was set as the ROI. This result was likely associated with the anatomical structure of the nose.

Noise Reduction Technique for Single-Color Video Plethysmography Using Singular Spectrum Analysis.

Recently, a contactless method for measuring a biological signal using a video camera has garnered attention. Especially, video plethysmography, a technique for obtaining a pulse wave from a video, is useful for managing the health of people on a daily basis. However, any body movement of a person subjected to the measurement leads to the generation of irregular noise in video plethysmography and reduces the accuracy of the recorded biological information, e.g., heart rate, during the measurement. Blind source separation is a popular technique for eliminating noise from the results of video plethysmography comprising different multiple-color channels. However, it is difficult to apply this technique to a single-color video such as a near-infrared video. Herein, a new method that combines singular spectrum analysis with the circular autocorrelation function is introduced to eliminate irregular noise in single-color video plethysmography. Applying the proposed method on videos collected from 39 individuals improved the estimation accuracy of instantaneous heart rate by approximately 44% over a conventional method using a linear filter. Furthermore, the proposed method also enabled more precise estimations of the heart rate than that achieved using multi-color video plethysmography.

Estimation of Absolute Blood Pressure Using Video Images Captured at Different Heights from the Heart.

The risk of cardiovascular diseases is related to the absolute level of blood pressure as well as its fluctuation while sleeping or during daily activities. To assess the risk, a simpler method to monitor daily blood pressure is desirable. In recent years, there has been a focus on developing a method to obtain pulse waves from video images of the human body. This is a promising technique to acquire biometric information without contact. In this study, we propose a new method to estimate the absolute level of blood pressure by using two video images of human hands captured at different heights from the heart. We focus on the amplitude difference of pulse waves obtained from the video images and derive an equation to estimate blood pressure based on the relationship between the internal pressure and the cross-sectional area of the blood vessel. The accuracy of the estimation was evaluated using data obtained from 5 healthy subjects performing cycling exercises that change their blood pressure. The average value of the root mean square error between the real value and the estimated value was 25.7 mmHg, while that of correlation coefficient was 0.66. There were large individual differences, particularly in the estimation of the absolute value of blood pressure. This result suggests the need for individual correction of the compliance curve, which represents the relationship between the internal pressure and the cross-sectional area of the blood vessel.

Sensorless cardiac phase detection for synchronized control of ventricular assist devices using nonlinear kernel regression model.

Recently, driving methods for synchronizing ventricular assist devices (VADs) with heart rhythm of patients suffering from severe heart failure have been receiving attention. Most of the conventional methods require implanting a sensor for measurement of a signal, such as electrocardiogram, to achieve synchronization. In general, implanting sensors into the cardiovascular system of the patients is undesirable in clinical situations. The objective of this study was to extract the heartbeat component without any additional sensors, and to synchronize the rotational speed of the VAD with this component. Although signals from the VAD such as the consumption current and the rotational speed are affected by heartbeat, these raw signals cannot be utilized directly in the heartbeat synchronization control methods because they are changed by not only the effect of heartbeat but also the change in the rotational speed itself. In this study, a nonlinear kernel regression model was adopted to estimate the instantaneous rotational speed from the raw signals. The heartbeat component was extracted by computing the estimation error of the model with parameters determined by using the signals when there was no effect of heartbeat. Validations were conducted on a mock circulatory system, and the heartbeat component was extracted well by the proposed method. Also, heartbeat synchronization control was achieved without any additional sensors in the test environment.

The Pilot Study of Evaluating Fluctuation in the Blood Flow Volume of the Radial Artery, a Site for Traditional Pulse Diagnosis.

BACKGROUND: Radial artery (RA) pulse diagnosis has been used in traditional Asian medicine. Blood pressure (BP) and pulse rate related to heart rate variability (HRV) can be monitored via the RA. The fluctuation in these parameters has been assessed using fast Fourier transform (FFT) analytical methods that calculate power spectra. METHODS: We measured blood flow volume (Volume) in the RA and evaluated its fluctuations. Normal participants (n = 34) were enrolled. We measured the hemodynamics of the right RA for approximately 50 s using ultrasonography. RESULTS: The parameters showed the center frequency (CF) of the power spectrum at low frequency (LF) and high frequency (HF). More than one spectral component indicated that there were fluctuations. The CF at LF for Volume was significantly different from that for vessel diameter (VD); however, it was significantly correlated with blood flow velocity (Velocity). On the other hand, the CF at HF for Volume was significantly different from that for Velocity; however, it was significantly correlated with VD. CONCLUSION: It is suggested that fluctuation in the Volume at LF of RA is influenced by the fluctuation in Velocity; on the other hand, fluctuation in the Volume at HF is influenced by the fluctuation in VD.

Blood perfusion display based on video pulse wave.

In this study, an easy system for monitoring dynamic blood perfusion patterns and the pulse wave velocity (PWV) has been developed by processing video images of a human body to assess blood circulation for daily management of physical conditions or for detecting persons in poor physical condition in public places. The experiment suggested that this tool can be used to easily evaluate the PWV; however, the obtained value from the video image of the face was about 1/10 of the standard value calculated from thick vessels. This difference may be related to the difference between thick vessels and thin-branched arterioles.

Estimation of heart rate variability using a compact radiofrequency motion sensor.

Physiological indices that reflect autonomic nervous activity are considered useful for monitoring peoples' health on a daily basis. A number of such indices are derived from heart rate variability, which is obtained by a radiofrequency (RF) motion sensor without making physical contact with the user's body. However, the bulkiness of RF motion sensors used in previous studies makes them unsuitable for home use. In this study, a new method to measure heart rate variability using a compact RF motion sensor that is sufficiently small to fit in a user's shirt pocket is proposed. To extract a heart rate related component from the sensor signal, an algorithm that optimizes a digital filter based on the power spectral density of the signal is proposed. The signals of the RF motion sensor were measured for 29 subjects during the resting state and their heart rate variability was estimated from the measured signals using the proposed method and a conventional method. A correlation coefficient between true heart rate and heart rate estimated from the proposed method was 0.69. Further, the experimental results showed the viability of the RF sensor for monitoring autonomic nervous activity. However, some improvements such as controlling the direction of sensing were necessary for stable measurement.

Techniques for estimating blood pressure variation using video images.

It is important to know about a sudden blood pressure change that occurs in everyday life and may pose a danger to human health. However, monitoring the blood pressure variation in daily life is difficult because a bulky and expensive sensor is needed to measure the blood pressure continuously. In this study, a new non-contact method is proposed to estimate the blood pressure variation using video images. In this method, the pulse propagation time difference or instantaneous phase difference is calculated between two pulse waves obtained from different parts of a subject's body captured by a video camera. The forehead, left cheek, and right hand are selected as regions to obtain pulse waves. Both the pulse propagation time difference and instantaneous phase difference were calculated from the video images of 20 healthy subjects performing the Valsalva maneuver. These indices are considered to have a negative correlation with the blood pressure variation because they approximate the pulse transit time obtained from a photoplethysmograph. However, the experimental results showed that the correlation coefficients between the blood pressure and the proposed indices were approximately 0.6 for the pulse wave obtained from the right hand. This result is considered to be due to the difference in the transmission depth into the skin between the green and infrared light used as light sources for the video image and conventional photoplethysmogram, respectively. In addition, the difference in the innervation of the face and hand may be related to the results.

A kernel-based method for markerless tumor tracking in kV fluoroscopic images.

Markerless tracking of respiration-induced tumor motion in kilo-voltage (kV) fluoroscopic image sequence is still a challenging task in real time image-guided radiation therapy (IGRT). Most of existing markerless tracking methods are based on a template matching technique or its extensions that are frequently sensitive to non-rigid tumor deformation and involve expensive computation. This paper presents a kernel-based method that is capable of tracking tumor motion in kV fluoroscopic image sequence with robust performance and low computational cost. The proposed tracking system consists of the following three steps. To enhance the contrast of kV fluoroscopic image, we firstly utilize a histogram equalization to transform the intensities of original images to a wider dynamical intensity range. A tumor target in the first frame is then represented by using a histogram-based feature vector. Subsequently, the target tracking is then formulated by maximizing a Bhattacharyya coefficient that measures the similarity between the tumor target and its candidates in the subsequent frames. The numerical solution for maximizing the Bhattacharyya coefficient is performed by a mean-shift algorithm. The proposed method was evaluated by using four clinical kV fluoroscopic image sequences. For comparison, we also implement four conventional template matching-based methods and compare their performance with our proposed method in terms of the tracking accuracy and computational cost. Experimental results demonstrated that the proposed method is superior to conventional template matching-based methods.

Discrimination ability and reproducibility of a new index reflecting autonomic nervous function based on pulsatile amplitude of photoplethysmography.

A new physiological index (µ(PA)) is proposed to represent the autonomic nervous system (ANS) function. The index µ(PA) is defined as the natural logarithm of the ratio between two different frequency components of the pulsatile amplitude of the photoplethysmogram (PPG) signal. The discrimination ability and the reproducibility of µ(PA) have been compared with other traditional ANS indices. In the experiment, the electrocardiogram, the PPG and continuous blood pressure were measured in 59 healthy young subjects (age 25.7 ± 6.3) and 86 healthy elderly subjects (age 70.2 ± 4.1) at rest. The discrimination ability and the reproducibility were evaluated by Cohen's d between young and elderly groups and by the interclass correlation coefficient, respectively. The results showed that the elderly subjects were significantly (p<;0.001) lower than young subjects in µ(PA) and a few traditional indices introduced to be compared with µ(PA). Therefore, it suggests that µ(PA) is associated with the decrease in the ANS function accompanied by aging. Moreover, it showed that the discrimination ability and the reproducibility of the proposed index are comparable or larger than those of traditional indices. The proposed index based on the PPG signal will be applied to tele-healthcare systems for monitoring people's health in daily life in combination with the ratio of the standard deviation of the R-R intervals to their average value (CVRR).

Evaluation of motor performances of hemiplegic patients using a virtual cycling wheelchair: an exploratory trial.

Cycling is known to be an effective rehabilitation exercise for hemiplegic patients who face difficulty during walking because of stroke or other brain disorders. A cycling wheelchair (CWC) is a useful tool to provide exercise for these patients and improve their quality of life. In previous studies, our group developed a system that allows patients to safely practice driving a CWC in a virtual environment. However, it has been difficult to check their motor performances and determine the effects of the exercise on a daily basis. This study is an exploratory trial for developing a method to evaluate the motor performances of users based on their CWC pedaling patterns. An experiment with some hemiplegic patients and healthy subjects was conducted and their pedaling patterns were analyzed. Results showed a significant difference between the hemiplegic patients and healthy subjects in an index that reflects pedaling balance between the feet. This result indicates a possible method of evaluating the motor performances of users based on their pedaling patterns.

Volume registration based on 3-D phase correlation for tumor motion estimation in 4-D CT.

This paper presents a three-dimensional (3-D) volume registration method that uses 3-D phase correlation to estimate the respiration-induced tumor motion in four-dimensional (4-D) thorax computed tomography (CT) for radiation therapy. The proposed method is an extension of 2-D phase correlation method to 3-D volume registration. Given two CT volumes obtained from different respiration stages, the tumor motion is modeled as a translational shift between the volumes. The 3-D phase correlation is obtained from the 3-D inverse Fourier transform of a normalized cross power spectrum of the volumes. The tumor motion along three directions is estimated by locating the highest peak in the 3-D phase correlation. In order to improve the estimation accuracy, we extend the 3-D phase correlation to sub-voxel accuracy. Experimental results demonstrate the effectiveness of the proposed method relative to a conventional 2-D phase correlation-based method.

Expansion capsules for diet control with artificial organ technology.

When we consider the medical economy, the Obesity is one of the leading preventable causes of death worldwide. However, a lot of previous scientific papers reported that 95% of obesity patients would not be able to control their weight by the diet. The surgical operation has been considered to the subjects with severe obesity. But, there is a possibility of complication or comorbidity in surgical operation. Tohoku University started to develop the expanding capsule with transcutaneous energy transmission system (TETS) having the same effect as the surgical operation. The capsule in the stomach will expand mechanically by energy transmission from outsides of the body, when the obesity patients will felt hungry. Small linear drive with folding umbrella type actuator would enable us the expansion of the diet capsules. Satisfactory characteristic of the energy transmission was obtained by the trial model of TETS during animal experiments. Animal experiments with healthy adult goats enabled us the evaluation of the inner stomach pressure time series changes, and feasibility study. Double blind test of the expanding capsule is now under planning. If the expanding capsule diet control system will be embodied, it becomes the gospel of the obese subject.

A time-varying seasonal autoregressive model-based prediction of respiratory motion for tumor following radiotherapy.

To achieve a better therapeutic effect and suppress side effects for lung cancer treatments, latency involved in current radiotherapy devices is aimed to be compensated for improving accuracy of continuous (not gating) irradiation to a respiratory moving tumor. A novel prediction method of lung tumor motion is developed for compensating the latency. An essential core of the method is to extract information valuable for the prediction, that is, the periodic nature inherent in respiratory motion. A seasonal autoregressive model useful to represent periodic motion has been extended to take into account the fluctuation of periodic nature in respiratory motion. The extended model estimates the fluctuation by using a correlation-based analysis for adaptation. The prediction performance of the proposed method was evaluated by using data sets of actual tumor motion and compared with those of the state-of-the-art methods. The proposed method demonstrated a high performance within submillimeter accuracy. That is, the average error of 1.0 s ahead predictions was 0.931 ± 0.055 mm. The accuracy achieved by the proposed method was the best among those by the others. The results suggest that the method can compensate the latency with sufficient accuracy for clinical use and contribute to improve the irradiation accuracy to the moving tumor.

Markerless Lung Tumor Motion Tracking by Dynamic Decomposition of X-Ray Image Intensity.

We propose a new markerless tracking technique of lung tumor motion by using an X-ray fluoroscopic image sequence for real-time image-guided radiation therapy (IGRT). A core innovation of the new technique is to extract a moving tumor intensity component from the fluoroscopic image intensity. The fluoroscopic intensity is the superimposition of intensity components of all the structures passed through by the X-ray. The tumor can then be extracted by decomposing the fluoroscopic intensity into the tumor intensity component and the others. The decomposition problem for more than two structures is ill posed, but it can be transformed into a well-posed one by temporally accumulating constraints that must be satisfied by the decomposed moving tumor component and the rest of the intensity components. The extracted tumor image can then be used to achieve accurate tumor motion tracking without implanted markers that are widely used in the current tracking techniques. The performance evaluation showed that the extraction error was sufficiently small and the extracted tumor tracking achieved a high and sufficient accuracy less than 1 mm for clinical datasets. These results clearly demonstrate the usefulness of the proposed method for markerless tumor motion tracking.