Nail oximeter for heart disease diagnosis and use method thereof

Abstract

The invention relates to the technical field of medical equipment, and discloses a nail oximeter for heart disease diagnosis and a using method thereof, the nail oximeter comprises a main body formed by pivoting an upper cover and a lower shell, and a U-shaped spring, a PCBA mainboard, a light-emitting tube and a photosensitive receiving unit are arranged in the main body. The method comprises the steps of collecting photoelectric volume pulse wave signals and segmenting the photoelectric volume pulse wave signals into cardiac cycle waveform sequences, calculating rhythm features and morphological distortion degree features for each waveform in combination with an average morphological template exclusive to a user, constructing multi-dimensional state vectors, mapping continuous vector sequences into geometric trajectories in a state space, and calculating the morphological distortion degree of each waveform according to the geometric trajectories. And finally, judging whether the heart rhythm state is sinus rhythm or abnormal rhythm by analyzing the macroscopic geometrical characteristics of the trajectory, and outputting a result. According to the method, the rhythm and form information of the heartbeat is fused for state space analysis, so that the accuracy and reliability of screening abnormal states such as arrhythmia in a household scene are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of medical device technology, specifically to a nail pulse oximeter for diagnosing heart diseases and its usage method. Background Technology

[0002] Finger-clip pulse oximeters, as a mature portable medical electronic device, rely on photoplethysmography (PPG) to non-invasively measure blood oxygen saturation and pulse rate from the fingertip. Due to their ease of operation and controllable cost, these devices have been widely used in medical monitoring and home health management.

[0003] With the increasing public demand for cardiovascular health management, the market is seeking solutions that allow for convenient screening of cardiac rhythm abnormalities in a home environment. To this end, some existing finger-clip pulse oximeters are beginning to integrate preliminary cardiac rhythm status analysis functions into their existing features.

[0004] Currently, the mainstream technology for these devices to achieve this function is based on statistical analysis of the time intervals between continuous pulse waves detected in the PPG signal, determining whether the heart rhythm is normal by judging the regularity of these time intervals. However, this analysis method, which relies solely on a single temporal rhythm dimension, has inherent limitations that are difficult to overcome in practical applications. In non-ideal measurement environments such as home use, factors such as user movement, unstable finger placement, or poor contact between the device and the skin can easily introduce signal noise and motion artifacts. These interferences can severely affect the accurate positioning of the pulse wave peak, leading to significant errors in pulse interval calculation and greatly reducing the reliability of the diagnostic results.

[0005] More importantly, this analytical method, which focuses solely on time intervals, completely ignores the wealth of physiological and pathological information inherent in the morphology of the photoplethysmography (PPG) waveform itself. The health of the heart's electrical and mechanical activity is reflected not only in the rhythm of the heartbeat but also in subtle features such as the rise and fall slopes, and the morphology of the main wave and dicrotic wave in each pulse wave. Because current technology fails to effectively integrate the rhythmic and morphological information of the pulse to form a multi-dimensional comprehensive evaluation system, its ability to screen for complex cardiac arrhythmias is fundamentally limited, making it difficult to meet users' needs for higher accuracy and reliability in health monitoring. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a nail pulse oximeter for diagnosing heart disease and its usage method. It solves the problems of existing arrhythmia diagnosis methods based on photoplethysmography (PPG) signals, which typically only focus on changes in the time interval of the heartbeat cycle and are not sensitive enough to morphological changes in the cardiac cycle waveform, and whose reference templates lack individualized adaptability.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A first aspect provides a nail pulse oximeter for diagnosing heart disease, comprising an upper cover and a lower cover, with two decorative covers rotatably connected between the upper and lower covers. A PCBA main board is fixedly connected inside the upper cover, a memory is fixedly connected to the top of the PCBA main board, and a processor is fixedly connected to the top of the PCBA main board. A face cover is fixedly connected to the top of the upper cover, and an LED lens is fixedly connected to the top of the face cover. An upper silicone pad is fixedly connected to the bottom of the upper cover, and a photosensitive receiving unit is fixedly connected to the outside of the upper silicone pad. A lower silicone pad is fixedly connected to the top of the lower cover, and a U-shaped spring is fixedly connected between the upper and lower silicone pads. A tail cover is fixedly connected to the outside of the lower cover, and a light-emitting tube is fixedly connected to the outside of the lower cover. A battery cover is provided at the bottom of the lower cover, and two batteries are installed at the bottom of the lower cover. The processor is configured to execute instructions stored in the memory to run a built-in cardiac rhythm diagnostic system, which processes photoplethysmography (PPG) signals acquired by the LED and the photosensitive receiving unit to determine the cardiac rhythm state.

[0008] Preferably, the cardiac rhythm diagnostic system includes: The signal processing module is used to segment the cardiac cycle waveform sequence from the photoplethysmography pulse wave signal; The template generation module is used to receive the cardiac cycle waveform sequence and generate a user-specific average morphology template based on a preset number of waveforms taken from the cardiac cycle waveform sequence. The feature construction module is used to process the cardiac cycle waveform sequence. By comparing each cardiac cycle waveform in the sequence with the user-specific average morphological template, a multi-dimensional state vector containing rhythm features and morphological distortion features is constructed for each cardiac cycle waveform, thereby forming a multi-dimensional state vector sequence corresponding to the cardiac cycle waveform sequence. A trajectory generation module is used to receive the multidimensional state vector sequence and convert the sequence into a corresponding trajectory. The state determination module is used to ultimately determine the heart rhythm state by analyzing the geometric features of the trajectory.

[0009] Preferably, the specific steps for generating the user-specific average shape template are as follows: The preset number of cardiac cycle waveforms are subjected to amplitude normalization, then the time axis is aligned using a dynamic time warping algorithm, and the aligned cardiac cycle waveforms are averaged point by point.

[0010] Preferably, the template generation module is further used for: When the determined heart rhythm state is a preset stable state, the user-specific average morphology template is updated by weighting the heart cycle waveform belonging to the stable state.

[0011] The update process can follow the formula below: ; in: This is an updated user-specific average pattern template; This is the user-specific average pattern template before the update; The waveform of a single cardiac cycle belonging to the said steady state; The preset weighted update coefficient has a value between 0 and 1.

[0012] Preferably, the multidimensional state vector further includes energy distribution characteristics obtained by time-frequency analysis of the cardiac cycle waveform.

[0013] Preferably, the state determination module is further used for: Calculate the difference between two adjacent multidimensional state vectors in the multidimensional state vector sequence to obtain a state transition vector; Based on the magnitude and direction of the state transition vector, the heart rhythm state is determined.

[0014] Preferably, the rhythmic feature is the time interval between the main peaks of two adjacent cardiac cycle waveforms.

[0015] Preferably, the value of the morphological distortion feature is the dynamic time-normalized distance between the cardiac cycle waveform and the user-specific average morphological template. For a length of User-specific average shape template and a length of The waveform of the cardiac cycle to be measured The dynamic time-warped distance between the two By building a The distance matrix is ​​calculated, and the elements in the matrix are... The calculation follows the following recursive relation: ; in: Points in the template sequence Points in the waveform to be measured The local distance cost between two data points is used to quantify the degree of dissimilarity between them. Indicates from three predecessor nodes , , Reach the current node The minimum cumulative distance of the three paths; This indicates that the first part of the template sequence will be... The points and the preceding waveform to be measured The minimum cumulative distance required to align points is obtained by adding the local distance cost of the current point to the aforementioned minimum cumulative distance.

[0016] Preferably, the specific steps for determining the cardiac rhythm state are as follows: The cardiac rhythm state is determined by analyzing the clustering, coverage volume, and periodicity of the trajectory.

[0017] A second aspect of the present invention provides a method for using a nail pulse oximeter for diagnosing heart disease, comprising the following steps: S1. Insert the user's finger between the upper and lower covers, so that the fingertip touches the lower silicone pad and the nail faces the upper silicone pad. S2. Start the nail pulse oximeter, thereby triggering the processor to control the light-emitting diode and photosensitive receiving unit to collect the photoplethysmography (PPG) signal, and segment the cardiac cycle waveform sequence from the PPG signal. S3. Based on a preset number of waveforms taken from the cardiac cycle waveform sequence, generate a user-specific average morphology template; S4. Process the cardiac cycle waveform sequence by comparing each cardiac cycle waveform in the sequence with the user-specific average morphological template to construct a multi-dimensional state vector containing rhythm features and morphological distortion features for each cardiac cycle waveform, thereby forming a multi-dimensional state vector sequence corresponding to the cardiac cycle waveform sequence. S5. Convert the multidimensional state vector sequence into a corresponding trajectory; S6. By analyzing the geometric features of the trajectory, the heart rhythm state is finally determined; S7. The determination result is output through the LED lens for reading.

[0018] This invention provides a nail pulse oximeter for diagnosing heart disease and its method of use. It has the following beneficial effects: 1. This invention overcomes the problem of diagnostic model incompatibility caused by individual physiological differences by generating a user-specific average morphological template as a comparison benchmark for subsequent cardiac cycle waveform morphology, thereby improving the accuracy of cardiac rhythm morphology analysis.

[0019] 2. This invention constructs a multi-dimensional state vector for each cardiac cycle waveform that simultaneously contains rhythmic features and morphological distortion features, thereby achieving synchronous quantification of a single heartbeat in both time and morphological dimensions, thus enabling a more comprehensive characterization of the heart rhythm state.

[0020] 3. This invention transforms the complex problem of judging heartbeat time series data into a geometric pattern recognition problem by converting multidimensional state vector sequences into geometric trajectories and analyzing their macroscopic geometric features. This enables a stable assessment of the overall state and dynamic trend of heart rhythm based on the clustering and periodicity of the trajectories. Attached Figure Description

[0021] Figure 1 This is a perspective view of the present invention; Figure 2 This is a schematic diagram of the PCBA motherboard of the present invention; Figure 3 This is a schematic diagram of the photosensitive receiving unit of the present invention; Figure 4 This is a schematic diagram of the U-shaped spring of the present invention; Figure 5 This is a schematic diagram of the light-emitting diode of the present invention; Figure 6 This is a schematic diagram of the battery of the present invention; Figure 7 This is a functional block diagram of the built-in cardiac rhythm diagnostic system in an embodiment of the present invention; Figure 8 A flowchart illustrating a method for using a nail pulse oximeter, as provided in an embodiment of the present invention.

[0022] The components include: 1. LED lens; 2. front cover; 3. PCBA main board; 4. decorative cover; 5. top cover; 6. upper silicone pad; 7. lower silicone pad; 8. tail cover; 9. U-shaped spring; 10. LED; 11. bottom cover; 12. battery; 13. battery cover; 14. processor; 15. memory; and 16. photosensitive receiving unit. Detailed Implementation

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see the appendix Figure 1 -Appendix Figure 7 This invention provides a nail pulse oximeter for diagnosing heart disease, including an upper cover 5 and a lower cover 11. Two decorative covers 4 are rotatably connected between the upper cover 5 and the lower cover 11. A PCBA motherboard 3 is fixedly connected inside the upper cover 5. A memory 15 is fixedly connected to the top of the PCBA motherboard 3. A processor 14 is fixedly connected to the top of the PCBA motherboard 3. A face cover 2 is fixedly connected to the top of the upper cover 5. An LED lens 1 is fixedly connected to the top of the face cover 2.

[0025] Specifically, the upper cover 5 and the lower cover 11 form the main frame of the device, protecting internal core components such as the PCBA motherboard 3 and battery 12 from external impacts or dust intrusion, extending the device's lifespan. Together, they create a gripping space suitable for finger insertion. The decorative cover 4 connects the upper cover 5 and the lower cover 11 and supports rotation, allowing adjustment of the gripping angle according to the user's finger size, improving user comfort and preventing discomfort caused by excessively tight or loose gripping. The PCBA motherboard 3, as the core hardware carrier, integrates key components such as the processor 14 and memory 15, enabling circuit connections and data communication between components, ensuring the overall coordination of the device's operation, and preventing operational failures caused by scattered component connections. The memory 15 stores program instructions, user-specific average morphology templates, raw signals, and intermediate data. The processor 14 executes the instructions stored in the memory 15, runs the built-in cardiac rhythm diagnostic system, and completes core data processing tasks such as signal segmentation, template generation, and feature construction. It is the core computing unit for the device to realize the function of diagnosing heart diseases. The cover 2 protects the top area of ​​the internal PCBA motherboard 3 and also serves as the mounting carrier for the display device, providing a stable installation environment for the display components. The LED lens 1 covers the display device, providing scratch and dust protection, preventing damage to the display device due to external factors, extending the life of the display device, reducing external light reflection, and improving the clarity of the displayed content.

[0026] The top cover 5 is fixedly connected to the bottom of the upper silicone pad 6, and the photosensitive receiving unit 16 is fixedly connected to the outside of the upper silicone pad 6. The bottom cover 11 is fixedly connected to the top of the lower silicone pad 7, and a U-shaped spring 9 is fixedly connected between the upper silicone pad 6 and the lower silicone pad 7. The bottom cover 11 is fixedly connected to the outside of the lower cover 11, and a light-emitting tube 10 is fixedly connected to the outside of the lower cover 11. The bottom of the lower cover 11 is provided with a battery cover 13, and two batteries 12 are installed at the bottom of the lower cover 11.

[0027] Specifically, the upper silicone pad 6 and the lower silicone pad 7 directly contact the user's fingers. The silicone material is soft and non-slip, which not only fixes the finger position but also improves wearing comfort and reduces discomfort during prolonged use. The upper silicone pad 6 is used to fix the photosensitive receiving unit 16, and the lower silicone pad 7 works in conjunction with the upper silicone pad 6 to hold the finger, indirectly ensuring the stability of signal acquisition and ensuring that the acquired signal meets the analysis standards. The U-shaped spring 9 provides elastic clamping force, which can adapt to fingers of different sizes, enhancing the device's compatibility with different users and eliminating the need for customized devices for different users. It maintains a stable distance between the silicone pads, avoiding the impact of excessive tightness on blood circulation due to pressure on the finger, or excessive looseness causing finger wobbling and unstable signal acquisition. The tail cover 8 covers the structural gaps at the edge of the lower cover 11 to prevent dust and moisture from entering and damaging components such as the light-emitting tube 10, ensuring safety. The normal operation of the core acquisition components optimizes the integrity of the device's appearance, enhances the product's quality, and strengthens user trust in the device. The light-emitting diode 10 and the photosensitive receiving unit 16 work together to acquire photoplethysmography (PPG) signals, providing raw data for cardiac rhythm diagnosis and forming the basis for all subsequent analysis. Their corresponding positions ensure the accuracy and stability of signal acquisition, reducing signal errors caused by positional deviations. The battery cover 13 covers the battery 12, facilitating user replacement of the battery 12 and preventing power outages due to loose or poor contact of the battery 12, ensuring stability during use and avoiding interruptions to the diagnostic process due to power failure. The battery 12 provides independent power to the device, eliminating the need for an external power source and ensuring portability. Users can use it anytime in various scenarios, such as at home or on the go, meeting their daily cardiac rhythm monitoring needs.

[0028] The processor 14 is configured to execute instructions stored in the memory 15 to run a built-in cardiac rhythm diagnostic system. The system processes photoplethysmography (PPG) signals acquired by the LED 10 and the photosensitive receiving unit 16 to determine the cardiac rhythm state.

[0029] Specifically, the processor 14 is a programmable computing device whose specific function is to acquire and execute program instructions stored in the memory 15, control the coordinated work of various hardware components of the device, and execute all data processing and calculation tasks of the cardiac rhythm diagnostic system, which will be detailed in subsequent chapters. The memory 15 is a computer-readable storage medium, which may be composed of non-volatile memory units and volatile memory units.

[0030] See attached document Figure 7 , Figure 7 This is a functional block diagram of the built-in cardiac rhythm diagnostic system in an embodiment of the present invention. The cardiac rhythm diagnostic system of the present invention is a set of program instructions stored in memory 15, loaded and executed by processor 14. Logically, the system includes: a signal processing module, a template generation module, a feature construction module, a trajectory generation module, and a state determination module.

[0031] In a diagnostic process, the initial input to the cardiac rhythm diagnostic system is the photoplethysmography (PPG) signal, which is acquired by the photosensitive receiving unit 16 and undergoes preliminary analog-to-digital conversion. This signal is first transmitted to the signal processing module.

[0032] The signal processing module performs segmentation processing on the received photoplethysmography (PPG) signal, resulting in a cardiac cycle waveform sequence containing multiple independent waveforms. The output of this cardiac cycle waveform sequence is simultaneously used as input to two subsequent modules: one part is sent to the template generation module, and the other part is sent to the feature construction module.

[0033] The template generation module receives a portion of the cardiac cycle waveform sequence and generates a user-specific average morphology template based on a specific algorithm. This generated user-specific average morphology template is then sent to the feature construction module as the second input data required for it to perform its feature construction task.

[0034] The feature construction module receives two inputs: a cardiac cycle waveform sequence and a user-specific average morphology template. Its function is to compare each cardiac cycle waveform in the sequence with the template, calculate and construct a multi-dimensional state vector for each waveform, thereby generating a multi-dimensional state vector sequence that corresponds one-to-one with the input waveform sequence.

[0035] The multidimensional state vector sequence is the input to the trajectory generation module. The trajectory generation module maps this vector sequence to a state space of a preset dimension and connects the vectors in the sequence in time order, thereby transforming and generating a geometric trajectory.

[0036] Finally, the geometric trajectory is transmitted to the state determination module. The state determination module performs quantitative analysis on the macroscopic geometric features of the trajectory and determines the final heart rhythm state based on the analysis results. This determined state is the final output of the entire system and can be transmitted to the display device on the faceplate 2 for display.

[0037] In a specific embodiment of the present invention, the function of the signal processing module is to receive continuous photoplethysmography pulse wave time series signals and extract a series of independent single cardiac cycle waveforms for subsequent analysis.

[0038] The raw PPG signal received by this module contains baseline drift caused by breathing and body movement, as well as high-frequency noise such as power line frequencies. To eliminate this interference, the signal processing module first applies a digital bandpass filter to the received signal. For example, a Butterworth bandpass filter with a passband range of 0.5 Hz to 8 Hz, which covers most of the frequency components of normal and abnormal heart rates.

[0039] After filtering, the signal processing module performs feature point detection on the filtered PPG signal to locate the boundaries of each cardiac cycle. Specifically, this is achieved by analyzing the local minima of the signal to identify the troughs at the beginning of each cardiac cycle. These troughs correspond temporally to the end of diastole and are relatively stable segmentation reference points.

[0040] The signal processing module defines the signal data segment between two consecutive troughs as an independent, complete single cardiac cycle waveform.

[0041] By sequentially performing this segmentation operation on the entire PPG time series signal, this module ultimately generates and outputs a cardiac cycle waveform sequence, denoted as... Each element Each segment is a waveform data segment representing a single heartbeat, consisting of a series of data points. This cardiac cycle waveform sequence is simultaneously transmitted to the template generation module and the feature construction module.

[0042] In a specific embodiment of the present invention, the template generation module functions to receive a cardiac cycle waveform sequence from the signal processing module and generate a user-specific average morphology template based thereon, which is used as a benchmark in subsequent analysis. This template is generated for the first time during the initial use of the device or according to a specific instruction.

[0043] Upon initial generation, the module selects a preset number of waveforms from the received cardiac cycle waveform sequence, for example, the initial 30 consecutive waveforms. To eliminate waveform duration differences caused by heart rate variations and amplitude differences caused by signal intensity fluctuations, the module performs a series of standardization processes on these 30 waveforms.

[0044] First, amplitude normalization is performed on the waveform for each cardiac cycle. Specifically, this is achieved using a min-max scaling method, which linearly maps the value of each waveform data point to an interval. This step eliminates amplitude inconsistencies between different cardiac cycles caused by variations in measurement conditions.

[0045] Next, time-axis alignment is performed on all amplitude-normalized cardiac cycle waveforms. Due to physiological fluctuations in heart rate, the lengths (i.e., the number of data points) of different cardiac cycle waveforms vary. This embodiment employs a dynamic time warping algorithm, using one waveform as a reference, to non-linearly scale the remaining waveforms along the time axis, ensuring that all waveforms are aligned at morphological feature points. This step ensures that the point-by-point correspondence in subsequent averaging calculations is based on morphological similarity, rather than simple temporal indexing.

[0046] After all selected cardiac cycle waveforms are aligned, the template generation module performs a point-by-point averaging operation. The amplitude values ​​of all aligned waveforms at each aligned time point are arithmetically averaged to obtain a single waveform representing the average shape of the user's heartbeat waveform over a specific period. This final generated waveform is the user-specific average shape template and is stored in the non-volatile storage unit of memory 15.

[0047] In addition, the template generation module is configured with a dynamic weighted update mechanism for the template. This update mechanism is triggered when the heart rhythm state output by the state determination module is a preset stable state (e.g., sinus rhythm). The module updates the stored user-specific average morphology template using the single cardiac cycle waveform currently determined to be in a stable state. The update process follows the recursive formula: ; In this formula: This represents the updated user-specific average pattern template; This represents the user-specific average pattern template before the update; This represents the current single cardiac cycle waveform, which has been confirmed to be in a stable state and has undergone the same normalization and alignment process. It is a preset weighted update coefficient, whose value ranges from 0 to 1.

[0048] This update mechanism allows the template to slowly adapt to long-term, gradual changes in the user's physiological state.

[0049] In a specific embodiment of the present invention, the function of the feature construction module is to receive the cardiac cycle waveform sequence from the signal processing module. And user-specific average shape templates from the template generation module. This module analyzes the waveform of each cardiac cycle in the sequence. Process it to construct a corresponding multidimensional state vector. Finally, it outputs a multidimensional state vector sequence corresponding to the input waveform sequence. .

[0050] Each multidimensional state vector It contains features with at least two dimensions: rhythmic features and morphological distortion features. Therefore, the vector The basic structure can be represented as .

[0051] Rhythmic characteristics The calculation method is as follows: determine the waveform of the current cardiac cycle. Main peak position and the waveform of the previous cardiac cycle Main peak position The main peak is the point of maximum amplitude within the waveform data segment. The rhythmic characteristic is the time interval between these two main peaks. This feature reflects the instantaneous rate of the heartbeat.

[0052] Morphological distortion characteristics The calculation method is as follows: quantify the waveform of the current cardiac cycle. User-specific average pattern template The morphological differences between them. In this embodiment, the value of this feature is defined as... and The dynamic time-normalized distance between them. Before calculation, first analyze the current cardiac cycle waveform. Perform the same magnitude normalization process as during template generation. The core of calculating this distance is filling in a cumulative distance matrix. Any element in the matrix The value follows the recursive formula: ; In this formula: Indicates that the template The former The subsequence composed of points and the waveform to be measured The former The minimum cumulative distance required to align a subsequence consisting of points; template The first in Data points With the waveform to be measured The first in Data points The local distance cost between them can be specifically calculated as the absolute difference in their magnitudes. Partially represents the journey from three adjacent predecessor nodes to the current node. The minimum cumulative distance among the three paths.

[0053] Finally, after the entire matrix is ​​filled, it will be located at the end of the matrix. value That is, two complete sequences and The DTW distance between them is used as a morphological distortion feature. ,in and They are templates and waveform The length.

[0054] In an optional embodiment, the multidimensional state vector It can also include a third dimension: energy distribution characteristics. This feature is achieved by analyzing the waveform of a single cardiac cycle. This is obtained by applying time-frequency analysis methods. By calculating the ratio of the energy of the waveform in a specific low-frequency band to the energy in a specific high-frequency band, a numerical value that quantifies the frequency structure change of the waveform can be obtained, serving as a characteristic of the energy distribution.

[0055] After processing all waveforms in the cardiac cycle waveform sequence, the feature construction module generates a multidimensional state vector sequence. Output to the trajectory generation module.

[0056] In a specific embodiment of the present invention, the trajectory generation module receives the output from the feature construction module, namely a complete multidimensional state vector sequence. The function of this module is to convert this numerical vector sequence into a geometric trajectory in a multidimensional state space, so as to facilitate subsequent geometric feature analysis.

[0057] The state space is a Cartesian coordinate system, and its number of dimensions is equal to that of each multidimensional state vector. The number of features contained in it. In one embodiment, if the state vector If only rhythmic features and morphological distortion features are included, then the state space is a two-dimensional plane. If the state vector... If it also includes energy distribution characteristics, then the state space is a three-dimensional space.

[0058] The transformation process performed by the trajectory generation module includes two steps. The first step involves transforming each multidimensional state vector in the sequence... It is plotted in the state space as an independent coordinate point. For example, a two-dimensional vector. Mapped as a point in the state space .

[0059] The second step is to map all the state points. They are connected according to their chronological order in the original sequence. Specifically, at point... With point A directed line segment is generated between the points, at the point... With point Generate a directed line segment between each point, and so on, until the last point is connected. .

[0060] This series of sequentially connected directed line segments together constitutes a complete geometric trajectory representing the evolution of the heart's rhythmic state during this period. This geometric trajectory is the final output of the trajectory generation module and is transmitted to the state determination module.

[0061] In a specific embodiment of the present invention, the state determination module receives the geometric trajectory generated by the trajectory generation module. The function of this module is to determine the final cardiac rhythm state by quantitatively analyzing the macroscopic geometric features of the trajectory. The determination logic of this module is based on the principle that different cardiac rhythm states will form trajectories with significantly different geometric features in the state space.

[0062] The state determination module primarily analyzes three geometric characteristics of the trajectory: clustering, coverage volume, and periodicity. For a trajectory composed of... State points The trajectory is constructed, and these characteristics are calculated as follows: The quantification of clustering is the average distance from all state points to their geometric center (centroid). The quantification of coverage volume is the volume of the smallest boundary hyperrectangle that can completely enclose all state points. The quantification of periodicity is the frequency with which the trajectory repeatedly traverses the same small volume region in the state space.

[0063] The state determination module internally stores a set of predefined thresholds used to map calculated geometric feature values ​​to specific cardiac rhythm states. For example, when the clustering index of the trajectory is less than the first threshold and its coverage volume index is less than the second threshold, the system determines the current cardiac rhythm state as sinus rhythm. This type of trajectory is a compact, clustered trajectory.

[0064] Conversely, when the clustering index of the trajectory exceeds the first threshold and its coverage volume index exceeds the second threshold, the system identifies the current cardiac rhythm as an abnormal rhythm, such as atrial fibrillation. This type of trajectory is diffuse and irregular.

[0065] In an optional embodiment, the state determination module is further configured with an auxiliary determination function based on state transition vectors. A state transition vector is defined as the vector difference between two consecutive state points in the state space. This module calculates the statistical variance of the magnitudes of all state transition vectors throughout the entire trajectory. A small variance value indicates that the heartbeat state transitions are stable and have consistent amplitudes, providing supplementary evidence for determining sinus rhythm. A large variance value indicates that the heartbeat state transitions are unstable and have drastic amplitude variations, providing supplementary evidence for determining abnormal heart rhythms.

[0066] The final output of this module is a label that identifies the specific heart rhythm state, which can be transmitted to a display device on the faceplate 2 for display.

[0067] See attached document Figure 1 and attached Figure 3 The present invention provides a method for using a fingernail pulse oximeter, the first step of which is device preparation and finger placement.

[0068] Before performing this step, the nail pulse oximeter should be in standby or off state. The user first applies a separating force to the upper cover 5 and the lower cover 11 to overcome the elastic restoring force of the U-shaped spring 9, thereby creating an opening between the upper silicone pad 6 and the lower silicone pad 7 large enough to accommodate the fingertip.

[0069] The user then inserts the distal knuckle of one finger into the inlet. The finger is placed in a specific orientation and position: the fingertip, the fleshy part containing abundant capillaries, is placed steadily on the surface of the lower silicone pad 7; at the same time, the nail side of the finger faces and contacts the surface of the upper silicone pad 6.

[0070] When the user removes the applied separation force, the elastic restoring force of the U-shaped spring 9 acts on the finger through the upper silicone pad 6 and the lower silicone pad 7, fixing the finger in place within the receiving cavity with a constant clamping force. This step establishes a precise physical alignment between the optical components inside the device and the user's finger tissue for signal acquisition.

[0071] This physical alignment is specifically manifested as follows: the light-emitting tube 10, located inside the lower cover 11, has its central axis aligned with the user's fingertip; while the photosensitive receiving unit 16, located inside the upper cover 5, has its photosensitive receiving surface aligned with the user's fingernail. This arrangement forms a transmissive light path, ensuring that the light emitted by the light-emitting tube 10 can penetrate the finger tissue and be effectively received by the photosensitive receiving unit 16. Simultaneously, the upper silicone pad 6 and lower silicone pad 7, in addition to fixing the finger, also serve to block external ambient light, thereby reducing interference with the photosensitive receiving unit 16.

[0072] See attached document Figure 8 This figure is a flowchart illustrating a method for using a fingernail pulse oximeter according to an embodiment of the present invention. After placing the finger in step S1, the user performs step S2, which activates the device and begins signal acquisition and segmentation. This step is triggered by pressing a start button located on the faceplate 2.

[0073] The pressing action closes a circuit, thereby waking up the processor 14. The awakened processor 14 begins executing preset program instructions stored in the memory 15. First, the processor 14 controls the LED 10 to emit light of a specific wavelength at a preset frequency and intensity. At the same time, the processor 14 activates the photosensitive receiving unit 16 to perform synchronous sampling.

[0074] The light emitted by the LED 10 penetrates the user's fingertip tissue and is received by the photosensitive receiving unit 16. Because blood volume pulsatiles with the heart cycle, the intensity of the received light also exhibits corresponding periodic fluctuations. The photosensitive receiving unit 16 converts this light signal into an analog electrical signal, which is then sampled and quantized by an analog-to-digital converter to form a raw digital photoplethysmography (FP-P) time-series signal.

[0075] The raw digital PPG signal is transmitted to processor 14, where processor 14 performs the functions of the signal processing module. First, a digital bandpass filter is applied to the signal, with its passband set to 0.5 Hz to 8 Hz, to filter out baseline drift caused by breathing and minor limb movements, as well as high-frequency noise introduced by the circuitry.

[0076] Specifically, processor 14 searches for and locates a series of local minima in the signal, which correspond to the troughs of each cardiac cycle. Then, processor 14 extracts and defines the data segment between any two consecutive troughs as an independent, complete single cardiac cycle waveform. .

[0077] By continuously segmenting the acquired PPG signal, step S2 ultimately generates and stores a cardiac cycle waveform sequence in the volatile storage unit of memory 15. This sequence is the direct input for subsequent steps S3 and S4.

[0078] See attached document Figure 7 and attached Figure 8 After generating and storing the cardiac cycle waveform sequence in step S2, the processor 14 automatically and continuously executes steps S3 to S6 to complete the internal processing and determination of the cardiac rhythm state. These steps are completed inside the device without user intervention.

[0079] Step S3 involves generating or loading a user-specific average pattern template. This step has two execution paths. In the first path, when the device is first used by the user or after a reset operation, the processor 14 performs the initial template generation. It extracts an initial, preset number (e.g., 30) waveforms from the waveform sequence generated in step S2, performs amplitude normalization and time axis alignment based on a dynamic time warping algorithm on each waveform. After processing, the processor 14 performs a point-by-point arithmetic mean on all aligned waveforms to generate the final user-specific average pattern template. And it is permanently stored in the non-volatile cell of memory 15. The second path is if a template already exists in the non-volatile memory cell. If so, processor 14 directly reads the template for use in subsequent steps.

[0080] Step S4 involves constructing a multidimensional state vector for each cardiac cycle waveform. Processor 14 iterates through the waveform sequence generated in step S2. For each cardiac cycle waveform in the sequence Processor 14 calculates its corresponding two-dimensional state vector. .

[0081] Among them, rhythmic characteristics The calculation method is as follows: Processor 14 in the current waveform and preceding waveform Locate the maximum amplitude points for each element and calculate the time interval between these two points. Morphological distortion characteristics. The calculation method is as follows: Processor 14 calculates the current waveform. (After amplitude normalization) and the template obtained in step S3 The dynamic time-warped distance between them. This distance is calculated by filling a cumulative distance matrix. To obtain matrix elements The calculation follows a recursive formula: ; in Template point With waveform points The local distance between them. Processor 14 combines the calculated two eigenvalues ​​into a vector. This process is repeated until corresponding state vectors have been generated for all waveforms in the sequence, forming a multidimensional state vector sequence. .

[0082] Step S5 generates a geometric trajectory from the vector sequence. Processor 14 processes the multidimensional state vector sequence generated in step S4. Mapped to a two-dimensional Cartesian coordinate system. Each vector in the sequence Considered as a point in this coordinate system Subsequently, processor 14 connects these points sequentially with directed line segments, following the original temporal order of the vectors in the sequence. This allows a complete geometric trajectory to be constructed within the volatile storage cells of memory 15.

[0083] Step S6 determines the cardiac rhythm state based on the geometric trajectory. Processor 14 quantitatively calculates the macroscopic geometric features of the geometric trajectory generated in step S5, mainly including the clustering of the trajectory point set (e.g., the average distance from all points to the centroid) and the coverage volume (e.g., the area of ​​the minimum bounding rectangle). Processor 14 compares the calculated geometric feature values ​​with a set of preset thresholds stored in non-volatile memory 15. If the trajectory exhibits strong clustering and a small coverage volume, it is determined to be sinus rhythm; if the trajectory exhibits diffuse, irregular, and large coverage volume, it is determined to be abnormal rhythm. This determination result, i.e., the specific cardiac rhythm state label, is the final output of this internal processing flow.

[0084] See attached document Figure 1 and attached Figure 8 After the internal processing flow in step S6 is completed, the processor 14 holds a cardiac rhythm status tag representing the final diagnostic result. At this time, the method enters the last step S7, which is the reading of the diagnostic result.

[0085] See attached document Figure 2 The processor 14 generates a specific set of display control commands based on the heart rhythm status label. These commands are sent via printed circuits on the PCBA motherboard 3 to the drive circuit of the display device integrated with the faceplate 2. In this embodiment, the display device is an organic light-emitting diode screen capable of displaying text, numbers, and graphic symbols.

[0086] When the heart rhythm state determined by step S6 is sinus rhythm, processor 14 controls the LED screen to display a set of visual information indicating a normal result. The specific display content is a static text, such as "normal heart rate," accompanied by a green circular icon. Simultaneously, the screen can also display information based on rhythm characteristics. The real-time heart rate value is calculated from the continuous value, for example, 72 beats / minute.

[0087] When the heart rhythm state determined by step S6 is an abnormal heart rhythm, the processor 14 controls the LED screen to display a set of warning visual information. A specific display content is a warning text, such as arrhythmia or suspected abnormality, accompanied by a warning symbol or color that distinguishes it from the normal state, such as a flashing orange triangle icon.

[0088] The user can read the final heart rhythm diagnosis result by directly observing the LED lens 1 on the faceplate 2. After displaying the result for a preset time (e.g., 30 seconds), or after the photosensitive receiving unit 16 detects that the finger has been removed and the signal is lost, the processor 14 will control the device to automatically enter a low-power standby state or shut down completely to save the power of the battery 12. The completion of this step marks the end of a complete heart rhythm diagnosis process.

Claims

1. A nail pulse oximeter for diagnosing heart disease, characterized in that, Includes an upper cover (5) and a lower cover (11). Two decorative covers (4) are rotatably connected between the upper cover (5) and the lower cover (11). A PCBA motherboard (3) is fixedly connected inside the upper cover (5). A memory (15) is fixedly connected to the top of the PCBA motherboard (3). A processor (14) is fixedly connected to the top of the PCBA motherboard (3). A face cover (2) is fixedly connected to the top of the upper cover (5). An LED lens (1) is fixedly connected to the top of the face cover (2). The bottom of the upper cover (5) is fixedly connected to... The upper silicone pad (6) is attached, and a photosensitive receiving unit (16) is provided on the outside of the upper silicone pad (6). The lower silicone pad (7) is fixedly connected to the top of the lower cover (11). A U-shaped spring (9) is fixedly connected between the upper silicone pad (6) and the lower silicone pad (7). A tail cover (8) is fixedly connected to the outside of the lower cover (11). A light-emitting tube (10) is fixedly connected to the outside of the lower cover (11). A battery cover (13) is provided at the bottom of the lower cover (11). Two batteries (12) are installed at the bottom of the lower cover (11). The processor (14) is configured to execute instructions stored in the memory (15) to run a built-in cardiac rhythm diagnostic system, which processes photoplethysmography signals collected by the light-emitting diode (10) and the photosensitive receiving unit (16) to determine the cardiac rhythm state.

2. The nail pulse oximeter for diagnosing heart disease according to claim 1, characterized in that, The cardiac rhythm diagnostic system includes: The signal processing module is used to segment the cardiac cycle waveform sequence from the photoplethysmography pulse wave signal; The template generation module is used to receive the cardiac cycle waveform sequence and generate a user-specific average morphology template based on a preset number of waveforms taken from the cardiac cycle waveform sequence. The feature construction module is used to process the cardiac cycle waveform sequence. By comparing each cardiac cycle waveform in the sequence with the user-specific average morphological template, a multi-dimensional state vector containing rhythm features and morphological distortion features is constructed for each cardiac cycle waveform, thereby forming a multi-dimensional state vector sequence corresponding to the cardiac cycle waveform sequence. A trajectory generation module is used to receive the multidimensional state vector sequence and convert the sequence into a corresponding trajectory. The state determination module is used to ultimately determine the heart rhythm state by analyzing the geometric features of the trajectory.

3. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The specific steps for generating the user-specific average morphology template are as follows: The preset number of cardiac cycle waveforms are subjected to amplitude normalization, then the time axis is aligned using a dynamic time warping algorithm, and the aligned cardiac cycle waveforms are averaged point by point.

4. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The template generation module is also used for: When the determined heart rhythm state is a preset stable state, the user-specific average morphology template is updated by weighting the heart cycle waveform belonging to the stable state.

5. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The multidimensional state vector also includes energy distribution characteristics obtained by time-frequency analysis of the cardiac cycle waveform.

6. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The state determination module is also used for: Calculate the difference between two adjacent multidimensional state vectors in the multidimensional state vector sequence to generate a state transition vector; Based on the magnitude and direction of the state transition vector, the heart rhythm state is determined.

7. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The rhythmic characteristic is the time interval between the main peaks of two adjacent cardiac cycle waveforms.

8. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The value of the morphological distortion feature is the dynamic time-normalized distance between the cardiac cycle waveform and the user-specific average morphological template.

9. A nail pulse oximeter for diagnosing heart disease according to claim 2, characterized in that, The specific steps for determining the cardiac rhythm state are as follows: The cardiac rhythm state is determined by analyzing the clustering, coverage volume, and periodicity of the trajectory.

10. A method of using a nail pulse oximeter for diagnosing heart disease, applied to the nail pulse oximeter for diagnosing heart disease as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. Insert the user's finger between the upper cover (5) and the lower cover (11), so that the fingertip contacts the lower silicone pad (7) and the nail faces the upper silicone pad (6). S2. Start the fingernail pulse oximeter, thereby triggering the processor (14) to control the light-emitting diode (10) and the photosensitive receiving unit (16) to collect the photoplethysmography (PPG) signal and extract the cardiac cycle waveform sequence from the PPG signal. S3. Based on a preset number of waveforms taken from the cardiac cycle waveform sequence, generate a user-specific average morphology template; S4. Process the cardiac cycle waveform sequence by comparing each cardiac cycle waveform in the sequence with the user-specific average morphological template to construct a multi-dimensional state vector containing rhythm features and morphological distortion features for each cardiac cycle waveform, thereby forming a multi-dimensional state vector sequence corresponding to the cardiac cycle waveform sequence. S5. Convert the multidimensional state vector sequence into a corresponding trajectory; S6. By analyzing the geometric features of the trajectory, the heart rhythm state is finally determined; S7. The determination result is output through the LED lens for reading.

Images