Accurate pulse wave velocity calculation by template matching

Abstract

The present invention relates to a method for determining a pulse wave velocity and / or a pulse transit time of a patient. The method comprises receiving at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient and determining a first seismocardiogram template, wherein the first seismocardiogram template is at least partially based on the at least two seismocardiograms. The method further comprises determining a time shift with respect to the first seismocardiogram template for each of the at least two seismocardiograms, shifting each of the at least two seismocardiograms based on the determined respective time shift and determining a second seismocardiogram template, wherein the second template is based on the at least two time-shifted seismocardiograms.

Description

[0001] Accurate Pulse Wave Velocity Calculation by Template Matching

[0002] Described herein are a method, an apparatus and a computer program for determining a pulse wave velocity and / or the pulse transit time of a patient, wherein determining the pulse wave velocity and / or the pulse transit time is at least partially based on at least two seismocardiograms of the patient.

[0003] Obtaining information about vital and / or biological parameters of a patient, e.g., blood pressure, pulse, breathing rate, oxygen saturation, heart rate variability and / or blood glucose becomes more and more popular in our daily lives and is nowadays ubiquitous. In particular, obtaining such information is not reserved for purely medical applications anymore but has also arrived in our private lives. Initially, obtaining information about vital and / or biological parameters was mainly used in hospitals and / or in the context of high-performance sports.

[0004] In hospitals, those parameters are used to monitor the health state of a patient before and / or after a treatment, e.g., a surgery, such that a documentation of the health state becomes available. In particular, the documentation of the health state can be used to recognize an improvement and / or a degradation of the health state of the patient. In high-performance sports, these vital and / or biological parameters are monitored and used to record the performance of the athlete in the context of the applied training method. In particular, this record may be used to identify customized training methods, which are optimal for the individual athlete. Based on the record, the training method may be adapted.

[0005] However, monitoring vital and / or biological parameters also becomes more important for home diagnostics and in the field of leisure sports, e.g., sports that is not performed at a professional level. For example, due to the overstressed health system, a patient is typically not subjected to an inpatient treatment but to an outpatient treatment, wherein the aftercare is performed remotely, e.g., at the home of the patient. For example, the patient is instructed to monitor after a treatment his pulse and / or his blood pressure, e.g., after a cardiac treatment.

[0006] In order to obtain this medical data a chest strap is typically used. The chest strap may comprise sensors, a processor and a memory. For example, the sensors of the chest strap directly record the heart beats at the chest of the patient and transmit the signals to the processor and / or to the memory, where the data is stored, e.g., for obtaining a cardiac monitoring for an extended period. Those data may then be read and analysed by medical staff, e.g., a medical doctor.

[0007] However, typically this data is subjected to temporal fluctuations and / or to measurement errors, which can significantly degrade the informative value of the data. In particular, with the emergence of more efficient smart phones which may comprise sensors capable of obtaining medical data of the user of said smart phone, these temporal fluctuations and / or measurement errors may become more problematic. An important parameter in this context is the pulse wave velocity and / or the pulse transit time, as the pulse wave velocity and / or the pulse transit time allows to draw conclusions about the blood pressure of the patient.

[0008] Therefore, there is a need for an improved method for determining the pulse wave velocity and / or the pulse transit time of the patient addressing at least some of the above-described disadvantages of the prior art and further improving other aspects.

[0009] A first aspect of the present invention relates to a method for determining a pulse wave velocity and / or the pulse transit time of a patient. The method comprises receiving at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient. The method further comprises determining a first seismocardiogram template, wherein the first seismocardiogram template is at least partially based on the at least two seismocardiograms and determining a time shift with respect to the first seismocardiogram template for each of the at least two seismocardiograms. Furthermore, the method comprises shifting each of the at least two seismocardiograms based on the determined respective time shift as well as determining a second seismocardiogram template, wherein the second seismocardiogram template is based on the at least two time-shifted seismocardiograms.

[0010] The method is directed to determining a pulse wave velocity and / or the pulse transit time of the patient. A pulse wave velocity may comprise a velocity at which the blood pressure pulse propagates through the circulatory system. Generally, the method may further comprise determining a pulse transit time. A pulse transit time may, for example, be the time that elapses between the opening of the aortic valve of the heart during a heartbeat and the arrival of the positive beat (in the sense of an increase in blood pressure) in a reference position, e.g., a fingertip.

[0011] In addition, or alternatively, the pulse wave velocity may be calculated from the pulse transit time, e.g., by taking into account a spatial distance between the heart and the reference position, e.g., the fingertip. In general, the relation between the pulse transit time and the pulse wave velocity may be given by the formula Pulse wave velocity = distance / pulse transit time.

[0012] The distance may be measured or estimated directly on the body using a measuring tape or similar.

[0013] A patient may be a person subjected to medical treatment, e.g., in a hospital and / or in a medical office. In addition, or alternatively, a patient may be any person aiming to obtain information of medical data, such as medical data that can be derived from the pulse wave velocity and / or the pulse transit time. For example, a patient can be any user of a device performing the method according to the present invention.

[0014] The method further comprises receiving at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient. A seismocardiogram may comprise a one-dimensional time curve of vibrations that can be measured on a surface. In addition, or alternatively, the seismocardiogram may comprise a two- and / or three-dimensional time curve of vibrations that can be measured on a surface. For example, the seismocardiogram may comprise a recording of cardiac vibrations that are measured on a human chest. The cardiac vibrations may be caused be a heartbeat of the patient. In particular, the vibrations may be caused by the contraction of heart muscles. Specially, each seismocardiogram at least partially comprises a heartbeat of the patient. For example, a first seismocardiogram may comprise a first heartbeat of the patient and a second seismocardiogram may comprise a second heartbeat of the patient. In particular, the first heartbeat and the second heartbeat may be consecutive heartbeats. In some embodiments, the seismocardiogram may comprise one heartbeat. In general, the first seismocardiogram may have a first duration and the second seismocardiogram may have a second duration. For example, the first duration and the second duration may be of equal length. In general, a seismocardiogram may comprise extreme points. For example, a seismocardiogram may comprise local maxima and / or local minima. In particular, a seismocardiogram may comprise local minima and local maxima. Specifically, the distribution of local minima and maxima may be characteristic for the heartbeat of the patient. The extreme points may differ for each heartbeat of the patient. For example, the first seismocardiogram associated with the first heartbeat may comprise a first set of extreme points and the second seismocardiogram associated with the second heartbeat may comprise a second set of extreme points.

[0015] The method comprises determining a first seismocardiogram template, wherein the first seismocardiogram template is at least partially based on the at least two seismocardiograms. In general, the first seismocardiogram template may comprise a seismocardiogram, e.g., the template may comprise a one-dimensional time curve of vibrations. In addition, or alternatively, the first seismocardiogram may comprise a two- and / or three-dimensional time curve of vibrations. Determining the first seismocardiogram template may comprise computing the first seismocardiogram template based at least partially on the at least two seismocardiograms. For example, computing the first seismocardiogram template may comprise computing the first seismocardiogram template based on a predefined function and / or algorithmic scheme. In particular, computing the first seismocardiogram template may comprise inserting the at least two seismocardiograms into the predefined function and / or applying the algorithmic scheme to the at least two seismocardiograms. In general, determining the first seismocardiogram template at least partially based on the at least two seismocardiograms may comprise determining the first seismocardiogram template based on data points of the at least two seismocardiograms and / or on datapoints derived from the at least two seismocardiograms. The first seismocardiogram template may comprise a set of data points. For example, the first seismocardiogram template may comprise a set of data points, wherein each data point associates a time point with an amplitude of the vibration. In addition, or alternatively, the first seismocardiogram template may comprise a function. Generally, the first seismocardiogram template may comprise extreme points, e.g., local minima and / or local maxima. For example, the extreme points of the first seismocardiogram template may be at least partially based on the extreme points of the first and / or the second seismocardiogram.

[0016] The method further comprises determining a time shift with respect to the first seismocardiogram template for each of the at least two seismocardiograms. In general, a time shift may comprise a period of time. The period of time may comprise negative and / or positive values. For example, a positive time shift value may indicate a time shift with respect to a right direction of a time axis. Similarly, a negative time shift value may indicate a time shift with respect to a left direction of the time axis. The absolute of the time shift value may indicate an amount of the time shift. For example, the absolute of the time shift value may indicate the amount of the time shift and the sign of the time shift value may indicate the direction of the time shift, e.g., in left and / or right direction of the time axis.

[0017] Generally, a time shift may comprise a period of time by which the seismocardiogram is shifted. For example, determining the time shift with respect to the first seismocardiogram template for a seismocardiogram may comprise determining a time shift for the seismocardiogram such that the timeshifted seismocardiogram and the first seismocardiogram template become more similar. Becoming more similar may comprise reducing a distance between the first seismocardiogram template and the seismocardiogram. For example, becoming more similar may comprise reducing a distance between at least a part of the set of extreme points of the first seismocardiogram template and at least a part of the extreme points of the seismocardiogram.

[0018] The method further comprises shifting each of the at least two seismocardiograms based on the determined respective time shift. For example, when the seismocardiogram comprises a set of points, each of the points may be time shifted. In particular, when each point is associated to a time and an amplitude of vibration, shifting the seismocardiogram may comprise to shift for each point the time value by the time shift. In addition, or alternatively, when the seismocardiogram comprises a function, shifting the seismocardiogram may comprise shifting the function with respect to the time axis. In particular, the function may be time shifted by the determined time shift. For example, shifting function may comprise a variable transformation of the function. Shifting each of the at least to seismocardiograms may result in at least two time-shifted seismocardiograms. In general, the time shift for the at least two seismocardiograms may differ. For example, there may be a first time shift for the first seismocardiogram and a second time shift for the second seismocardiogram.

[0019] The method further comprises determining a second seismocardiogram template, wherein the second seismocardiogram template is based on the at least two time-shifted seismocardiograms. The at least two time-shifted seismocardiograms may comprise the at least two seismocardiograms that have been shifted based on the determined respective time shift. In general, determining the second seismocardiogram template may be based on the same function and / or algorithmic scheme as determining the first seismocardiogram template. Alternatively, determining the second seismocardiogram template may be based on a different function and / or algorithmic scheme as determining the first seismocardiogram template.

[0020] Generally, the first and / or second seismocardiogram template may at least partially comprise a heartbeat of the patient. The heartbeat of the patient may not be a real heartbeat of the patient but may comprise a virtual and / or derived heartbeat of the patient. In other words, the first and / or second seismocardiogram template may comprise extreme points that are characteristic for a heartbeat of a patient, although the extreme points are not corresponding to a real heartbeat of the patient, e.g., a heartbeat of the patient that has actually been measured.

[0021] Determining the second seismocardiogram template according to the above-described method may result in an amplified seismocardiogram of the patient. More precisely, the at least two seismocardiograms may be subjected to temporal noise and / or to temporal fluctuations. For example, the blood pressure and thereby the pulse wave velocity and / or the pulse transit time vary from heartbeat to heartbeat, which leads to time shifts in the corresponding seismocardiograms. The time shifts may be typically in a range between 5 ms to 10 ms. However, the exact value of these time shifts for each heartbeat are unknown, leading to corresponding fluctuations in the determined pulse wave velocity and / or pulse transit time.

[0022] Therefore, determining the second seismocardiogram template based on the at least two time-shifted seismocardiograms, wherein the at least two time-shifted seismocardiograms are shifted such as to compensate for the temporal fluctuations, leads to an amplified seismocardiogram of the patient. The amplified seismocardiogram of the patient at least partially compensates for the unknown time shifts of the heartbeat. Therefore, the second seismocardiogram template may constitute an improved and corrected model of the seismocardiogram of the patient and thereby contributes to an improved determination of the pulse wave velocity and / or the pulse transit time. More precisely, the second seismocardiogram template minimizes the fluctuations in the at least two seismocardiograms, maximizes the quality of the data and thus optimizes the determination of the pulse wave velocity and / or the pulse transit time of the patient.

[0023] Generally, the method may further comprise determining the pulse wave velocity and / or the pulse transit time at least partially based on the second seismocardiogram template. In particular, determining the pulse wave velocity and / or the pulse transit time may be at least partially based on at least one local extreme point of the second seismocardiogram template.

[0024] The pulse wave velocity and / or the pulse transit time may be determined at least partially based on the second seismocardiogram template. For example, the pulse wave velocity and / or the pulse transit time may be determined at least partially based on the second seismocardiogram template such as to treat the second template as the actual seismocardiogram of the patient. In particular, at least one local extreme point of the second seismocardiogram template may be treated as at least one local extreme point of the actual seismocardiogram of the patient.

[0025] In some embodiments, determining the pulse wave velocity and / or the pulse transmit time may be further based on a photoplethysmogram of the patient. For example, determining the pulse wave velocity and / or the pulse transit time may be based on at least one characteristic point of a seismocardiogram and at least one characteristic point of a photoplethysmogram. In general, the at least one characteristic point of the seismocardiogram may be associated with a heartbeat of the patient and / or the at least one characteristic point of the photoplethysmogram may be associated with a heartbeat of the patient. Specifically, the at least one characteristic point of the seismocardiogram and the at least one characteristic point of the photoplethysmogram may be associated with the same heartbeat of the patient. For example, a first characteristic point of the seismocardiogram and a first characteristic point of the photoplethysmogram may be associated with a first heartbeat and a second characteristic point of the seismocardiogram and a second characteristic point of the photoplethysmogram may be associated with a second heartbeat of the patient. For example, the characteristic point may comprise an inflection point, a minimal turning point, a maximal turning point and / or a saddle point. In particular, the type of the characteristic point of the seismocardiogram may coincide with the type of characteristic point of the photoplethysmogram. For example, the characteristic point of the seismocardiogram and the characteristic point of the photoplethysmogram may comprise a maximal turning point. In addition, or alternatively, the type of the characteristic points may differ. For example, the characteristic point of the seismocardiogram may comprise a maximal turning point and the characteristic point of the photoplethysmogram may comprise an inflection point and / or a minimal turning point. Generally, the pulse wave velocity and / or the pulse transit time may be at least partially based on an offset between the seismocardiogram and the photoplethysmogram. The offset may comprise a delay or retardation between the seismocardiogram and the photoplethysmogram. For example, the delay between the seismocardiogram and the photoplethysmogram may comprise a delay and / or time shift between a characteristic point of the seismocardiogram and a characteristic point of the photoplethysmogram. In particular, the delay may comprise a delay between a characteristic point of the seismocardiogram and a characteristic point of the photoplethysmogram, wherein preferably the characteristic point of the seismocardiogram and the characteristic point of the photoplethysmogram are associated with the same heartbeat of the patient.

[0026] Generally, determining the pulse wave velocity and / or the pulse transit time may serve as an indicator for an arterial stiffness. For example, a higher pulse transit time / a lower pulse wave velocity may indicate a lower arterial stiffness. Similarly, a lower pulse transit time / a higher pulse wave velocity may indicate a higher arterial stiffness. In addition, or alternatively, a higher pulse transit time / a lower pulse wave velocity may indicate a lower blood pressure.

[0027] The second seismocardiogram template may constitute an improved and corrected model of the seismocardiogram of the patient and thereby contributes to an improved determination of the pulse wave velocity and / or the pulse transit time. More precisely, the second seismocardiogram template minimizes the fluctuations in the at least two seismocardiograms, maximizes the quality of the data and thus optimizes the determination of the pulse wave velocity and / or the pulse transit time of the patient.

[0028] In general, receiving the at least two seismocardiograms may comprise generating at least two excerpts of a seismocardiogram.

[0029] In general, the at least two seismocardiograms may be excerpts of a seismocardiogram. For example, the seismocardiogram may be a seismocardiogram of a patient that was measured over an extended period of time. For example, the seismocardiogram may comprise a continuous signal and / or a continuous delivery of data points over an extended period of time. For example, the seismocardiogram may be a seismocardiogram of the patient comprising a period of time of at least 60 s. In particular, the seismocardiogram may comprise a contiguous period of time. The seismocardiogram may comprise at least 60 heartbeats, preferably at least 70 heartbeats, most preferably at least 80 heartbeats. Generating at least two excerpts of the seismocardiogram may comprise to select at least two excerpts of the seismocardiogram. For example, the first excerpt of the seismocardiogram may be associated with a first period and corresponding data points from of the seismocardiogram and the second excerpt may be associated with a second period and corresponding data points from of the seismocardiogram. Data points from the seismocardiogram may comprise an amplitude. In some embodiments, the first period and the second period may be disjoint, e.g., the first and the second period essentially do not overlap. Generally, the first and the second period may be of the same duration. For example, the duration of the first and / or the second period may be chosen such that the corresponding seismocardiogram comprises at least one heartbeat. In particular, the duration of the first and / or second period may be chosen such that it comprises one heartbeat. In general, generating the at least two excerpts of the seismocardiogram may comprise treating the first period and corresponding data points from the seismocardiogram as a first seismocardiogram and treating the second period and corresponding data points from the seismocardiogram as a second seismocardiogram. In other words, generating at least two excerpts may comprise treating the at least two excerpts as separate and / or independent seismocardiograms.

[0030] Generating at least two excerpts from the seismocardiogram allows to obtain at least two seismocardiograms from the seismocardiogram. More precisely, a seismocardiogram measured over a period of time of 60 s may comprise at least 60 heartbeats. Therefore, at least 60 excerpts may be generated based on a seismocardiogram that has been recorded over a period of time of 60 s. Generally, generating at least two excerpts from the seismocardiogram may allow to obtain a high number of seismocardiograms in a relatively short time, e.g., within a reasonable time for the patient. Hence, generating excerpts contributes to the comfort of the patient when determining the pulse wave velocity and / or the pulse transit time. Furthermore, generating excerpts leads to a high number of seismocardiograms that may be used for determining the first and / or the second seismocardiogram template, thereby contributing to the quality of the second seismocardiogram template. In particular, a high number of seismocardiograms minimizes the fluctuations in the at least two seismocardiograms, maximizes the quality of the data and thus optimizes the determination of the pulse wave velocity and / or the pulse transit time of the patient.

[0031] Generally, each of the at least two seismocardiograms may be associated with a photoplethysmogram of the patient. Specifically, each of the at least two seismocardiograms may comprise a period which is associated with a respective peak of the photoplethysmogram.

[0032] The at least two seismocardiograms may be associated with a photoplethysmogram of the patient. A photoplethysmogram may comprise an optically obtained plethy smogram. A photoplethysmogram may be a one-dimensional time curve of colour change values of a skin surface that occur when an incoming pulse wave dilates the arterial vessels beneath the skin, so that more blood and thus more red blood cells are present.

[0033] For example, being associated with a photoplethysmogram may comprise that the photoplethysmogram and the seismocardiogram are associated with the same period. For example, the photoplethysmogram and the seismocardiogram may have been measured simultaneously. In this case, there may be a one-to-one correspondence between the periods associated with the at least two seismocardiograms and corresponding periods of the photoplethysmogram.

[0034] For example, if the at least two seismocardiograms are based on generating excerpts of a seismocardiogram, the seismocardiogram and the photoplethysmogram may have been measured simultaneously. In some embodiments, the photoplethysmogram and the seismocardiogram may have mean measured based on the same device. Generally, generating the excerpts of the seismocardiogram may be at least partially based on the photoplethysmogram. In particular, generating excerpts may be based on characteristic points and / or local extreme points of the photoplethysmogram. For example, the first period of the first excerpt and / or the second period of the second excerpt may be based on the photoplethysmogram. In particular, the period may be associated with a corresponding period of the photoplethysmogram. For example, the period of the photoplethysmogram may comprise a peak. The peak may comprise an early systolic peak and / or a late systolic peak and / or a dicrotic notch and / or a local minimum. In addition, or alternatively, the peak may comprise a point in a vicinity of an inflection point of the photoplethysmogram.

[0035] In particular, the end point of the period may be at least partially based on the respective associated peak of the photoplethysmogram. In addition, or alternatively, the length of the period may be at least 0.1 s, preferably at least 0.5 s, most preferably at least 1 s. In addition, or alternatively, the length of each period may be at most 3 s, preferably at most 2 s, most preferably at most 1.5 s.

[0036] The end point of the period may be at least partially based on the respective associated peak of the photoplethysmogram. For example, the end point of the period may be the point in time of the respective associated peak of the photoplethysmogram. In particular, when the seismocardiogram and the photoplethysmogram are measured simultaneously and / or refer to a common time frame, the period of the first seismocardiogram may be chosen such that the end point of the seismocardiogram corresponds to the point in time of the respective peak in the photoplethysmogram.

[0037] Generally, the length of the period may be at least 0.1s, preferably at least 0.5 s, most preferably at least 1 s. In other words, each of the at least two seismocardiograms may comprise a period of length at least 0.1s, preferably at least 0.5 s, most preferably at least 1 s. In particular, each of the at least two seismocardiograms may comprise data points associated with aperiod of length at least 0.1s, preferably at least 0.5 s, most preferably at least 1 s. For example, if the end point of the period is at least partially based on the respective associated peak of the photoplethysmogram, the period may extend from the point of time of the respective associated peak by at least 0.1s, preferably at least 0.5 s, most preferably at least 1 s.

[0038] In addition, or alternatively, the length of the period may be at most 3 s, preferably at most 2 s, most preferably at most 1.5 s. In other words, each of the at least two seismocardiograms may comprise a period of length at most 3 s, preferably at most 2 s, most preferably at most 1.5 s. In particular, each of the at least two seismocardiograms may comprise data points associated with aperiod of length at most 3 s, preferably at most 2 s, most preferably at most 1.5 s. For example, if the end point of the period is at least partially based on the respective associated peak of the photoplethysmogram, the period may extend from the point of time of the respective associated peak by at most 3 s, preferably at most 2 s, most preferably at most 1.5 s.

[0039] Associating each of the at least two seismocardiograms with a photoplethysmogram of the patient allows to synchronize the at least two seismocardiograms with the photoplethysmogram. In particular, associating the at least two seismocardiograms with the photoplethysmogram may allow to generate the at least two seismocardiograms with respect to a reference point. For example, using as a reference point a peak of the photoplethysmogram may allow for an accurate synchronization. An accurate synchronization allows for an accurate determination of the pulse wave velocity and / or the pulse transit time. By limiting the length of the period from below and / or above it can be ensured that the period comprises sufficient information about the seismocardiogram of the patient while at the same time reducing the amount of data that must been stored.

[0040] In general, the photoplethysmogram may be at least partially based on a photosensor. For example, the photosensor may comprise a camera. In particular, the camera may be comprised in a portable electronic device.

[0041] The photosensor may be designed to detect light emitted by a light source, i.e., it may detect the corresponding wavelength ranges by means of one or more photodiodes or similar. The photosensor may be designed for visible light. It may be designed as a smartphone camera, for example. The light signal may be preferably a digital signal. If a simple photodiode is used, a temporal progression of the light intensity may be recorded directly. This analog signal (usually a voltage curve over time) may be converted into a digital light signal using an analog -to-digital converter, for example. The light source may be a light-emitting diode (LED) that emits white light in the visible wavelength range between 380 nm and 780 nm wavelength. For example, a permanently switched-on LED may be used for the flash of a smartphone or tablet.

[0042] Alternatively, infrared light with a wavelength above 780 nm to 3 pm may be used in combination with red visible light (640 nm to 780 nm wavelength), as is common with pulse oximeters, to determine the oxygen saturation of the blood based on the different absorption properties of light for oxidized and non-oxidized haemoglobin. In principle, the wavelength range used may be matched to an absorption maximum of oxidized and / or non-oxidized haemoglobin. Traditional pulse oximeters may use two light sources for this purpose: a red light-emitting diode for a wavelength of 660 nm and an infrared light-emitting diode for a wavelength of 940 nm.

[0043] Similarly, the at least two seismocardiograms may be at least partially based on a measurement of an accelerometer and / or a gyroscopic sensor. Specifically, the accelerometer and / or the gyroscopic sensor may be comprised in a portable electronic device.

[0044] The measurement of the accelerometer and / or the gyroscopic sensor may comprise a measurement in at least one, preferably at least two, most preferably in at least three spatial directions.

[0045] Generally, when using a smartphone or a tablet, a measurement may be performed in such a way that a lying user places the device on the chest, e.g., positioning the device vertically on an area of skin above the sternum or above the heart at the 4th intercostal space, while placing a finger on the camera and the flash of the device. Whether the position on the sternum or above the heart at the 4th intercostal space is more suitable may be determined by carrying out a series of measurements for a user at both positions, e.g., lasting one minute.

[0046] For example, if the at least two seismocardiograms are based on generating at least two excerpts of a seismocardiogram of the patient, the seismocardiogram may be at least partially based on a measurement of an accelerometer. In some embodiments, generating the at least two excerpts may be performed in real-time with measuring the seismocardiogram of the patient partially based on the accelerometer.

[0047] In general, determining the time-shift may comprise minimizing a distance between each of the at least two seismocardiograms and the first seismocardiogram template.

[0048] The time-shift may be determined such that a distance between the at least time-shifted seismocardiograms and the first seismocardiogram template is minimized. For example, there may be a first distance between the first seismocardiogram and the first seismocardiogram template and a second distance between the second seismocardiogram and the first seismocardiogram template. The first time-shift associated with the first seismocardiogram may be chosen such that the distance between the corresponding time-shifted first seismocardiogram, e.g., the first seismocardiogram where the time-shift has been applied, and the first seismocardiogram template is smaller and / or equal to the first distance. Similarly, the second time-shift associated with the second seismocardiogram may be chosen such that the distance between the corresponding time-shifted second seismocardiogram, e.g., the second seismocardiogram where the time-shift has been applied, and the first seismocardiogram template is smaller and / or equal to the second distance. The first distance and the second distance may generally be different.

[0049] The distance may comprise a mathematical metric and / or norm. For example, each of the at least two seismocardiograms may comprise a set of data points, e.g., a set of data points associating a time value to an amplitude. In this case, the distance may comprise a discrete metric and / or norm, e.g., a metric and / or norm involving only a finite number of points. For example, the distance may comprise a sum over a plurality of distances.

[0050] In addition, or alternatively, the distance may comprise an evaluation of an integral. For example, if one of the at least two seismocardiograms is given as a function and / or the first seismocardiogram template is given as a function, the distance may comprise a function metric and / or function norm. In particular, the metric and / or norm may comprise a Lp metric and / or Lpnorm.

[0051] Determining the time-shift such that a distance between each of the at least two seismocardiograms and the first seismocardiogram template is minimized allows for the computation of the respective optimal time-shifts. Computing the respective optimal time-shifts may result in optimal time-shifted seismocardiograms and thus in an optimal second seismocardiogram template. Thereby, the temporal fluctuations in the at least two seismocardiograms can be optimally mitigated, minimizing the error in the determined pulse wave velocity and / or pulse transit time, thus contributing to an optimal determination of those values.

[0052] Generally, determining the first seismocardiogram template may be at least partially based on an average of the at least two seismocardiograms. In addition, or alternatively, determining the second seismocardiogram template may be at least partially based on an average of the at least two time-shifted seismocardiograms.

[0053] The first seismocardiogram template may be at least partially based on an average of the at least two seismocardiograms. For example, the average may comprise a mean value and / or a median and / or a function of the at least two seismocardiograms. In particular, the first seismocardiogram template may be at least partially based on an arithmetic mean of the at least to seismocardiograms. For example, the data points of the first seismocardiogram template may be based on an average of the data points of the at least two seismocardiograms. In particular, each data point of the first seismocardiogram template may be based on an average of corresponding data points of the at least two seismocardiograms. In addition, or alternatively, if the at least two seismocardiograms comprise a function, the first seismocardiogram template may be an average of the corresponding functions. In this case, the first seismocardiogram template may comprise a function.

[0054] Similarly, determining the second seismocardiogram template may be at least partially based on an average of the at least two time-shifted seismocardiograms. In general, the average (the method for averaging) used for the first seismocardiogram template may be the same as the average used for the second seismocardiogram template. In some embodiments, the method used for the first seismocardiogram template may differ from the method used for the second seismocardiogram template. For example, the first seismocardiogram template and the second seismocardiogram template may be determined at least partially based on a mean value, e.g., an arithmetic mean.

[0055] Determining the first seismocardiogram template and / or second seismocardiogram template at least partially based on an average of the at least two seismocardiograms allows for a minimization of the temporal fluctuations, e.g., the unknown time shifts, of the at least two seismocardiograms. More precisely, using an average may be optimal in order to minimize and / or mitigate fluctuations of statistical nature. Minimizing the fluctuations maximizes the quality of the data and thus optimizes the determination of the pulse wave velocity and / or the pulse transit time of the patient.

[0056] The at least two seismocardiograms may be based on a resampling, preferably wherein the resampling is based on a Hermite interpolation.

[0057] The at least two seismocardiograms may be based on a resampling. For example, the resampling may adjust and / or change a sampling rate of the at least two seismocardiograms. In particular, the at least two seismocardiograms may be based on a resampling such as to obtain a common time resolution of the at least two seismocardiograms. For example, the first seismocardiogram may initially comprise a first time resolution and the second seismocardiogram may initially comprise a second time revolution. Resampling the first and / or the second seismocardiogram may comprise that the resampled first and / or second seismocardiogram comprise the same time resolution. The resampling may be based on a Hermite interpolation of the at least two seismocardiograms. In addition, or alternatively, other interpolation methods such as linear interpolation and / or Lagrange interpolation and / or spline interpolation may be used. Using Hermite interpolation for the resampling of the at least two seismocardiograms allows for an efficient and precise computation of the interpolated function. In particular, using Hermite interpolation may allow to achieve a sub-sample precision, e.g., a precision that is higher than the original sampling rate. Thereby, a resampling, particularly a resampling based on Hermite interpolation, ensures a common time resolution of the at least two seismocardiograms and thus enhances the quality of the first and / or the second seismocardiogram template. In some embodiments, the resampling may in addition comprise a resampling of the photoplethy smogram. For example, the resampling may be such that the sampling rate of the photoplethy smogram and the at least two seismocardiograms essentially coincides.

[0058] Specifically, the resampling may comprise an upsampling of the at least two seismocardiograms to at least 600 Hz, preferably at least 800 Hz, most preferably at least 1000 Hz.

[0059] The resampling, preferably based on Hermite interpolation, may comprise an upsampling of the at least two seismocardiograms. For example, the at least two seismocardiograms may be upsampled to a sampling rate of at least 600 Hz, preferably at least 800 Hz, most preferably at least 1000 Hz. Generally, the sampling rate of the upsampling may at least partially depend on a sampling rate of the photoplethy smogram. In particular, the upsamling may be such that the sampling rate of the at least two seismocardiogram and the photoplethysmogram essentially coincide. In addition, or alternatively, the sampling rate of the upsampling may be at least partially based on a maximal sampling rate of the photoplethysmogram and / or the at least two seismocardiograms. In some embodiments, the photoplethysmogram and the at least two seismocardiograms may be upsampled.

[0060] In general, determining the pulse wave velocity and / or the pulse transit time of the patient may comprise iteratively determining a sequence of seismocardiogram templates, wherein each seismocardiogram template is based on the preceding seismocardiogram template.

[0061] The determination of the pulse wave of the patient may comprise iteratively determining a sequence of seismocardiogram templates. In particular, each seismocardiogram template may be based on a preceding seismocardiogram template. For example, based on the at least two seismocardiograms, the first seismocardiogram template may be determined. Subsequently, based on the first seismocardiogram template and the at least two seismocardiograms at least two time-shifted seismocardiograms may be computed. Based on the at least two time-shifted seismocardiograms the second seismocardiogram template may be determined. Iteratively determining a sequence of seismocardiogram templates may comprise that the determined second seismocardiogram template may be treated as a fist seismocardiogram template. More precisely, upon determining the second seismocardiogram template, iteratively determining may comprise determining a time shift with respect to the second seismocardiogram template for the at least two seismocardiograms and shifting the at least two seismocardiograms based on the determined respective time shift (based on the second template). Iteratively determining may further comprise determining a third seismocardiogram template, wherein the third seismocardiogram template is based on the at least two time-shifted seismocardiograms. Generally, this process may be iterated, e.g., by now treating the third seismocardiogram template as the first seismocardiogram template.

[0062] Iteratively determining a sequence of seismocardiogram templates may improve the quality of the seismocardiogram template with each iteration. For example, the temporal fluctuations of a seismocardiogram template may be reduced with each iteration. Generally, reducing the temporal fluctuations may be associated with a decrease in the determined time shifts for each iteration. In particular, while the peaks and / or local extrema within the first seismocardiogram template may be washed out due to the fluctuations, with each iteration the peaks and / or local extrema may become more explicit and sharper. Thereby, the seismocardiogram associated with the seismocardiogram template will be amplified. Thus, iteratively determining a sequence of seismocardiogram templates minimizes the temporal fluctuations, maximizes the quality of the seismocardiogram template and hence optimizes the determination of the pulse wave velocity and / or the pulse transit time of the patient.

[0063] Specifically, iteratively determining the sequence of seismocardiogram templates may end based on a convergence of the iterated seismocardiogram templates. For example, iteratively determining the sequence may end based on a difference of two seismocardiogram templates being smaller than a threshold.

[0064] The iterative determination of the sequence of seismocardiogram templates may end based on a convergence of the iterated seismocardiogram templates. A convergence of the iterated seismocardiogram templates may comprise that a distance and / or difference between two seismocardiogram templates becomes smaller than a threshold. For example, the distance and / or difference may be computed between two consecutive iteratives. The threshold may be a predefined threshold. In some embodiments, where the method is performed by a portable electronic device, such as a mobile phone and / or a tablet computer, the threshold may be defined by the user, e.g., the patient. The distance and / or the difference may be determined by the same method used for the determination of the time shift. For example, the distance and / or difference may be based on a metric and / or norm. If the iteratively determining ends, the obtained (latest) seismocardiogram template may be used for determining the pulse wave velocity and / or the pulse transit time. A further aspect of the present invention relates to a computer program. The computer program comprises instructions which, when the program is executed by a computer, cause the computer carry out the above-described method.

[0065] A further aspect of the present invention relates to an apparatus for determining a pulse wave velocity and / or a pulse transit time of a patient. The apparatus comprises means for receiving at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient and means for determining a first seismocardiogram template, wherein the first seismocardiogram template is at least partially based on the at least two seismocardiograms. The apparatus further comprises means for determining a time shift with respect to the first seismocardiogram template for each of the at least two seismocardiograms and means for shifting each of the at least two seismocardiograms based on the determined respective time shift. The apparatus further comprises means for determining a second seismocardiogram template, wherein the second template is based on the at least two time-shifted seismocardiograms.

[0066] Generally, the apparatus may further comprise means for determining the pulse wave velocity at least partially based on the second seismocardiogram template, preferably on at least one local extreme point of the second seismocardiogram template. In addition, or alternatively, the apparatus may further comprise means for determining the pulse transit time at least partially based on the second seismocardiogram template, preferably on at least one local extreme point of the second seismocardiogram template.

[0067] Furthermore, the apparatus may comprise means for receiving at least one photoplethysmogram of the patient.

[0068] In general, the apparatus may be configured to perform at least partially the above-described method for determining a pulse wave velocity and / or a pulse transit time of a patient.

[0069] In some embodiments, the means for receiving at least two seismocardiograms of the patient may comprise a motion detector. In particular, the motion detector may comprise an accelerometer or a gyroscopic sensor. Generally, the motion detector and / or the accelerometer and / or the gyroscopic sensor may be comprised in a portable electronic device. For example, the portable electronic device may comprise a mobile phone and / or a tablet computer. The motion detector and / or the accelerometer and / or the gyroscopic sensor may be such as to detect movements in at least one dimension, preferably at least two dimensions, most preferably at least three dimensions. The means for receiving at least one photoplethysmogram of the patient may comprise a photosensor. For example, the photosensor may comprise a camera. In addition, or alternatively, the means for receiving at least one photoplethysmogram may comprise a light source. For example, the light source may comprise at least one light-emitting diode. The light-emitting diode may emit light with a wavelength in the range from 380 nm to 780 nm. In addition, or alternatively, the light source may emit light with a wavelength in a range from 780 nm to 3 pm, preferably in combination with red visible light, e.g., light with a wavelength in the range from 640 nm to 780 nm. Generally, the photosensor, particularly the camera of a portable electronic device, may be configured to capture light emitted of the light source.

[0070] The means for receiving at least one photoplethysmogram may be at least partially comprised in a portable electronic device. The portable electronic device may comprise a mobile phone and / or a tablet computer and / or a smart watch. In particular, the light source may comprise a light-emitting diode of the portable electronic device, specifically a diode configured for a flash of the camera of the portable electronic device. For a typical mobile phone and / or tablet computer the camera of the mobile phone and / or tablet computer may be positioned in a vicinity of the light-emitting diode of the mobile phone and / or light-emitting diode.

[0071] In some embodiments, the means for receiving at least two seismocardiograms of the patient and the means for receiving at least one photoplethysmogram may be both comprised in a portable electronic device, e.g., a mobile phone and / or a tablet computer. For example, the mobile phone and / or the portable electronic device may comprise a camera, at least one light-emitting diode for the flash of the camera and an accelerometer.

[0072] For example, if a portable electronic device, e.g., mobile phone and / or a tablet computer is used, determining the pulse wave velocity and / or the pulse transit time may comprise to position the portable electronic device in a region of the patient associated with the sternum of the patient. For example, the portable electronic device may be positions on the sternum of the patient. In particular, the portable electronic device may be positioned essentially perpendicular to the region of the patient. In addition, or alternatively, the portable electronic device may be positioned essentially perpendicular to a region of the patient associated with a fourth intercostal space of the patient.

[0073] In addition, or alternatively, the determining the pulse wave velocity and / or the pulse transit time may comprise to position a finger of the patient above the photosensor and / or the light source. For example, the finger of the patient may be positioned such as to at least partially cover the photosensor and the light source. In particular, the finger may at least partially cover the camera and the light source of the portable electronic device, e.g., a mobile phone and / or a tablet computer. Whether aspects are implemented as hardware or software means depends upon the particular application and design constraints imposed on the overall system. By way of example, an element, or any portion of an element, or any combination of elements may be implemented as a processing system that may include one or more processors. Examples of processors include microprocessors, microcontrollers, graphics processing units (GPUs), central processing units (CPUs), application processors, digital signal processors (DSPs), reduced instruction set computing (RISC) processors, systems on a chip (SoC), baseband processors, field programmable gate arrays (FPGAs), programmable logic devices (PEDs), state machines, gated logic, discrete hardware circuits, and other suitable hardware configured to perform the various functionality described throughout this disclosure. One or more processors in the processing system may execute software. Software shall be construed broadly to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software components, applications, software applications, software packages, routines, subroutines, objects, executables, threads of execution, procedures, functions, etc., whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise.

[0074] Accordingly, in one or more exemplary embodiments, the functions described may be implemented in hardware, software, or any combination thereof. If implemented in software, the functions may be stored on or encoded as one or more instructions or code on a computer-readable medium. Computer- readable media includes computer storage media. Storage media may be any available media that can be accessed by a computer. By way of example, and not limitation, such computer-readable media can comprise a random-access memory (RAM), a read-only memory (ROM), an electrically erasable programmable ROM (EEPROM), optical disk storage, magnetic disk storage, other magnetic storage devices, combinations of the aforementioned types of computer-readable media, or any other medium that can be used to store computer executable code in the form of instructions or data structures that can be accessed by a computer.

[0075] It is noted that any combination of features that have been described above as belonging to certain embodiments / aspects of the present invention is also an embodiment of the present invention, provided such a feature combination is feasible, i.e., does not lead to any contradictions.

[0076] In the following, exemplary embodiments of the invention are described with reference to the figures.

[0077] The figures show:

[0078] Fig. 1 schematic representation of a first embodiment of a method for determining the pulse wave velocity and / or the pulse transit time of a patient; Fig. 2 schematic representation of a second embodiment of a method for determining the pulse wave velocity and / or the pulse transit time of a patient;

[0079] Fig. 3 exemplary illustration of a seismocardiogram, a photoplethysmogram and generating at least two seismocardiograms based on the seismocardiogram;

[0080] Fig. 4 exemplary illustration of upsampling a seismocardiogram and a photoplethysmogram to a common sampling rate;

[0081] Fig. 5 exemplary illustration of a system for obtaining a seismocardiogram and a photoplethysmogram .

[0082] Figure 1 is a schematic representation of a method 100 for determining a pulse wave velocity and / or the pulse transit time of a patient. The method 100 comprises receiving 110 at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient. The method 100 further comprises determining 120 a first seismocardiogram template, wherein the first template is at least partially based on the at least two seismocardiograms as well as determining 130 a time shift with respect to the template for each of the at least two seismocardiograms. The method further comprises shifting 140 each of the at least two seismocardiograms based on the determined respective time shift and determining 150 a second seismocardiogram template, wherein the second template is based on the at least two time-shifted seismocardiograms. In some embodiments, the method further comprises determining 160 the pulse wave velocity and / or the pulse transit time at least partially based on the second template.

[0083] Generally, the method 100 for determining the pulse wave velocity and / or the pulse transit time of the patient may comprise determining a pulse transit time of the patient. In particular, determining 160 the pulse wave velocity and / or the pulse transit time at least partially based on the second template may comprise determining a pulse transit time of the patient. For example, the method 100 may comprise determining a pulse transit time and deriving at least partially based on the pulse transit time the pulse wave velocity. In addition, or alternatively, the method 100 may comprise determining the pulse wave velocity and deriving at least partially based on the pulse wave velocity a pulse transit time.

[0084] Receiving 110 at least two seismocardiograms of the patient may comprise generating at least two excerpts of a seismocardiogram (see e.g., Fig. 3). For example, each of the at least two excerpts may be associated with a heartbeat of the patient, e.g., with two consecutive heartbeats of the patient. In some embodiments, the first excerpt may be associated with the first seismocardiogram and the second excerpt may be associated with the second seismocardiogram. For example, the seismocardiogram may be a seismocardiogram comprising at least two contiguous heartbeats of the patient. In particular, the seismocardiogram may comprise at least 20, preferably at least 40, more preferably at least 60, most preferably at least 80 heartbeats of the patient.

[0085] In general, the at least two seismocardiograms may be at least partially based on a measurement of an accelerometer and / or a gyroscopic sensor. For example, the accelerometer may be at least partially comprised in a portable electronic device. The portable electronic device may comprise a mobile phone and / or a tablet computer. In some embodiments, e.g., wherein receiving 110 at least two seismocardiograms comprises generating at least two excerpts of a seismocardiogram, the seismocardiogram may be at least partially based on a measurement of an accelerometer.

[0086] Generally, each of the at least two seismocardiograms may be associated with a photoplethysmogram of the patient. In particular, each of the at least two seismocardiograms may comprise a period which is associated with a respective peak of the photoplethysmogram. For example, generating the at least two excerpts of the seismocardiogram may be at least partially based on the associated photoplethysmogram. In particular, generating the at least two excerpts may be at least partially based on a respective peak of the photoplethysmogram.

[0087] Specifically, the end point of the period may be at least partially based on the respective associated peak of the photoplethysmogram. In addition, or alternatively, the length of the period may be at least 0.1 s, preferably at least 0.5 s, most preferably at least 1 s. In addition, or alternatively, the length of each period may be at most 3 s, preferably at most 2 s, most preferably at most 1.5 s.

[0088] Generally, the photoplethysmogram may be at least partially based on a photosensor. For example, the photoplethysmogram may be based on a photosensor comprised in a camera, Particular, the photosensor and / or the camera may be at least partially comprised in a portable electronic device. The portable electronic device may comprise a mobile phone and / or a tablet computer and / or a smart watch.

[0089] In general, determining 120 the first seismocardiogram template may comprise determining the first seismocardiogram template at least partially based on an average of the at least two seismocardiograms. In addition, or alternatively, determining 120 the second seismocardiogram template may be at least partially based on an average of the at least two time-shifted seismocardiograms. For example, determining 120 the first and / or the second seismocardiogram template at least partially based on an average may comprise determining a mean value and / or a median value of the at least two seismocardiograms and / or the at least two time-shifted seismocardiograms. Generally, determining 130 the time shift with respect to the first template for each of the at least two seismocardiograms may comprise minimizing a distance between each of the at least two seismocardiograms and the first template. For example, for each of the at least two seismocardiograms an individual time shift may be determined. In particular, a first time shift may be determined for the first seismocardiogram and a second time shift may be determined for the second seismocardiogram. Specifically, the first time shift and / or the second time shift may be such that the distance between the first template and the time-shifted first seismocardiogram (which has been shifted by the first time shift) and / or the distance between the first template and the time-shifted second seismocardiogram (which has been shifted by the second time shift) is minimized.

[0090] In general, shifting 140 each of the at least two seismocardiograms based on the determined respective time shift may comprise shifting the at least two seismocardiograms with respect to a time axis. For example, the time shift may comprise negative and / or positive values. In particular, a negative value may indicate a shift towards the left direction of the time axis and a positive value may indicate a shift towards the right direction of the time axis. In some embodiments, the association between positive / negative values and right / left direction may be vice versa. The amount of the shift may be at least partially determined by the absolute of the time shift.

[0091] Determining 150 the second seismocardiogram template may be at least partially based on an average of the at least two time-shifted seismocardiograms.

[0092] Figure 2 is a schematic illustration of a method 200 for determining the pulse wave velocity and / or the pulse transit time of a patient. In general, method 200 may be performed in combination with method 100. In particular, some steps and / or aspects of method 100 may replace some steps / 200 of the method 200.

[0093] The method 200 comprises receiving 210 at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient. The receiving 210 may at least partially comprise the step 110 of method 100.

[0094] The method 200 further comprises determining 220 a first seismocardiogram template, wherein the first template is at least partially based on the at least two seismocardiograms as well as determining 230 a time shift with respect to the template for each of the at least two seismocardiograms. The determining 220 and / or the determining 230 may at least partially comprise the step 120 and / or 130 of the method 100. - 1 -

[0095] The method 200 further comprises shifting 240 each of the at least two seismocardiograms based on the determined respective time shift and determining 250 a second seismocardiogram template, wherein the second template is based on the at least two time-shifted seismocardiograms. The shifting 240 and / or the determining 250 may at least partially comprise the steps 140 and / or 150 of method 100.

[0096] The method 200 further comprises iteratively determining 260 a sequence of templates, wherein each template is based on the preceding template. For example, after determining 250 the second seismocardiogram template, the iteratively determining 260 may comprise to return to determining 230 a time shift. In particular, the iteratively determining 260 may comprise to treat the second seismocardiogram template as a first seismocardiogram template and return to determining 230 a time shift. In other words, iteratively determining 260 may comprise determining 230 a time shift with respect to the first template for each of the at least two seismocardiograms, wherein the first template is identified with the second template.

[0097] Generally, the iteratively determining 260 may end based on a convergence of the iterated templates. For example, the iteratively determining 260 may end based on a difference between two templates being smaller than a threshold.

[0098] Optionally, the method 200 further comprises determining 270 the pulse wave velocity and / or the pulse transit time at least partially based on the second template. For example, determining 270 the pulse wave velocity and / or the pulse transit time may be performed after the iteratively determining 260. In particular, the determining 270 may be performed when the iteratively determining 260 has ended, preferably based on a difference between two templates being smaller than a threshold. For example, the determining 270 may be based on the last iterated template of the iteratively determining 260.

[0099] Figure 3 is exemplary illustration 300 of a seismocardiogram, a photoplethysmogram and generating at least two seismocardiograms based on the seismocardiogram. The illustration 300 shows a seismocardiogram 310 and a photoplethysmogram 320. The photoplethysmogram 320 comprises a plurality of peaks and / or local extrema 322, 324, 326, 330a, 330b. For example, the photoplethysmogram 320 comprises a systolic peak 322, a dicrotic notch 324 and a diastolic peak 326. The photoplethysmogram 320 further comprises inflection points 330a, 330b. In some embodiments, the inflection points 330a, 330b may comprise a point in a vicinity of the inflection point and / or a point at half distance between a local minimum of the photoplethysmogram 320 and the systolic peak 322. In addition, or alternatively, the inflection points maybe comprise a minimum of the photoplethysmogram . The seismocardiogram 310 comprises a plurality of heartbeats. For example, the seismocardiogram may comprise at least four, preferably at least six heartbeats. Similarly, the photoplethysmogram 320 comprises a plurality of heartbeats. For example, the photoplethysmogram may comprise at least four, preferably at least six heartbeats. In particular, each heartbeat may be associated with a respective systolic peak 322 and / or a respective dicrotic notch 324 and / or a respective diastolic peak 326 and / or a respective inflection point 330a, 330b. Generally, the seismocardiogram 310 and the photoplethysmogram 320 may be associated. For example, the seismocardiogram 310 and the photoplethysmogram 320 may share a common time frame, e.g., the seismocardiogram 310 and the photoplethysmogram 320 may have been recorded in the same time interval.

[0100] Generally, receiving 110, 210 of at least two seismocardiograms may comprise generating at least two excerpts 350a, 350b of the seismocardiogram 310. Generating the at least two excerpts 350a, 350b may be based on the associated photoplethysmogram 320. For example, generating the at least two excerpts 350a, 350b may be based on the inflection points 330a, 330b respectively. For example, the excerpt 350a may comprise a period 340a, wherein the end point 344a of the period 340a is based on the inflection point 330a. In particular, the end point 344a may coincide with the inflection point 330a. Similarly, the excerpt 350b may comprise a period 340b, wherein the end point 344b of the period 340b is based on the inflection point 330b. In particular, the end point 344b may coincide with the inflection point 330b. In addition, or alternatively, the length of the period 340a, 340b may be at least 0. 1 s, preferably at least 0.5 s, most preferably at least 1 s. For example, the length of the period 340a may be the difference between the end point 344a and the start point 342a of the period 340a. Similarly, the length of the period 340b may be the difference between the end point 344b and the start point 342b of the period 340b. In addition, or alternatively, the length of each period 340a, 340b may be at most 3 s, preferably at most 2 s, most preferably at most 1.5 s. Generally, the length of the period 340a may differ from the length of the period 340b. In some embodiments, the length of the period 340a essentially coincides with the length ofthe period 340b. Generally, the at least two excerpts 350a, 350b may be treated as seismocardiograms, e.g., the first seismocardiogram may be identified with the excerpt 350a and the second seismocardiogram may be identified with the excerpt 350b.

[0101] Figure 4 is an exemplary illustration of upsampling 400 a seismocardiogram 410 and a photoplethysmogram 420 to a common sampling rate. Generally, the photoplethysmogram 420 may comprise in a given period of time less data points than the seismocardiogram 410. For example, the photoplethysmogram 420 may comprise three data points 425a, 425b, 425c and the seismocardiogram may comprise eight data points 415a to 415h. The seismocardiogram 410 and / or the photoplethysmogram 420 may comprise an interpolation, e.g., the seismocardiogram 410 and / or the photoplethysmogram 420 may have been interpolated based on the data points 425a, 425b, 425c and 415a to 415h, respectively. In some embodiments the interpolation ofthe seismocardiogram 410 and / or the photoplethysmogram 420 may be based on a Hermite interpolation. Generally, based on the interpolation 410, 420, the seismocardiogram 410 and the photoplethysmogram 420 may be upsampled to a common sampling rate. For example, the seismocardiogram 410 and the photoplethysmogram 420 may be upsampled t a common sampling rate of at least 600 Hz, preferably at least 800 Hz, most preferably at least 1000 Hz. For example, the common sampling rate may comprise sampling the interpolated functions 410, 420 at equidistant time points 435a to 435b.

[0102] Figure 5 is exemplary illustration of a system 500 for obtaining a seismocardiogram and a photoplethysmogram. The system 500 may comprise a mobile phone and / or a tablet computer 510. The mobile phone and / or tablet computer 510 may comprise a photosensor 520 and an accelerometer 550. In some embodiments, the photosensor 520 may comprise a camera of the mobile phone and / or tablet computer 510. With respect to the system 500, the accelerometer 550 may be positioned at a bottom end of the mobile phone and / or tablet computer 510. However, in some embodiments, the accelerometer 550 may be located a different position within the mobile phone and / or tablet computer 510. The system further comprises a hand 540 of the patient and a finger 530 of the patient. Furthermore, the system comprises an area 560 associated with a sternum of the patient. In some embodiments, the area 560 may be associated with an area of the patient, wherein vibrations caused by the heartbeat of the patient are measurable for the accelerometer 550. Generally, determining the pulse wave velocity and / or the pulse transit time of the patient may comprise that the patient positions the mobile phone and / or tablet computer 510 within the area 560 and positions the finger 530 of the hand 530 at least partially over the photosensor 520.

Claims

1. Claims1. Method (100, 200) for determining a pulse wave velocity and / or a pulse transit time of a patient, comprising: receiving (110) at least two seismocardiograms of the patient, wherein each seismocardiogram at least partially comprises a heartbeat of the patient; determining (120) a first seismocardiogram template, wherein the first seismocardiogram template is at least partially based on the at least two seismocardiograms; determining (130) a time shift with respect to the first seismocardiogram template for each of the at least two seismocardiograms; shifting (140) each of the at least two seismocardiograms based on the determined respective time shift; and determining (150) a second seismocardiogram template, wherein the second template is based on the at least two time-shifted seismocardiograms.

2. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to claim 1, further comprising: determining (160) the pulse wave velocity at least partially based on the second seismocardiogram template, preferably on at least one local extreme point of the second seismocardiogram template; and / or determining (160) the pulse transit time at least partially based on the second seismocardiogram template, preferably on at least one local extreme point of the second seismocardiogram template.

3. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to claim 1 or 2, wherein receiving (110) the at least two seismocardiograms comprises generating at least two excerpts (350a, 350b) of a seismocardiogram (310).

4. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to one of the preceding claims, wherein each of the at least two seismocardiograms is associated with a photoplethysmogram (320) of the patient.

5. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to claim 4, wherein each of the at least two seismocardiograms comprises a period (340a, 340b) which is associated with a respective peak (330a, 330b) of the photoplethysmogram (320).

6. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to claim 5, wherein: an end point (3244a, 344b) of the period (340a, 340b) is at least partially based on the respective associated peak (330a, 330b) of the photoplethysmogram (320); and / or a length of the period (340a, 340b) is at least 0. 1 s, preferably at least 0.5 s, most preferably at least 1 s; and / or the length of each period (340a, 340b) is at most 3 s, preferably at most 2 s, most preferably at most 1.5 s.

7. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to one of the claims 4 to 6, wherein the photoplethysmogram (320) is at least partially based on a photosensor (520), preferably wherein the photosensor (520) comprises a camera, most preferably wherein the camera is comprised in a portable electronic device (510).

8. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to one of the preceding claims, wherein determining (130) the time-shift comprises minimizing a distance between each of the at least two seismocardiograms and the first seismocardiogram template.

9. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to one of the preceding claims, wherein: determining (120) the first seismocardiogram template is at least partially based on an average of the at least two seismocardiograms; and / or determining (150) the second seismocardiogram template is at least partially based on an average of the at least two time-shifted seismocardiograms.

10. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to one of the preceding claims, wherein the at least two seismocardiograms are based on a resampling (400), preferably wherein the resampling is based on a Hermite interpolation (410, 420).

11. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to claim 10, wherein the resampling (400) comprises an upsampling of the at least two seismocardiograms to at least 600 Hz, preferably at least 800 Hz, most preferably at least 1000 Hz.

12. Method (200) for determining the pulse wave velocity and / or the pulse transit time according to one of the preceding claims, wherein the method comprises iteratively determining (260) a sequence of seismocardiogram templates, wherein each seismocardiogram template is based on the preceding seismocardiogram template.

13. Method (200) for determining the pulse wave velocity and / or the pulse transit time according to claim 12, wherein the iteratively determining (260) the sequence of seismocardiogram templates ends based on a convergence of the iterated seismocardiogram templates, preferably wherein iteratively determining the sequence ends based on a difference of two seismocardiogram templates being smaller than a threshold.

14. Method (100, 200) for determining the pulse wave velocity and / or the pulse transit time according to any of the preceding claims, wherein the at least two seismocardiograms are at least partially based on a measurement of an accelerometer (550), preferably wherein the accelerometer is comprised in a portable electronic device (510).

15. A computer program comprising instructions which, when the program is executed by a computer, cause the computer carry out the method (100, 200) according to one of the claims 1 to 14.

Images