Method for contactless identification of a person

Abstract

A method for the contactless identification of a person on the basis of functional parameters of the cardiovascular system relates to the processing of medical diagnostic data. The technical result of the claimed invention is more accurate identification of a person, which is achieved using a method for the contactless identification of a person on the basis of functional parameters of the cardiovascular system characterized in that it entails recording a training set of functional parameters of the cardiovascular system, which is compiled by registering electrophysiological signals received simultaneously from an electrocardiogram, and by registering wireless communication channel state parameters, and training a machine learning model which in turn makes it possible to carry out identification of a person.

Description

[0001] METHOD OF CONTACTLESS IDENTIFICATION OF A PERSON

[0002] DESCRIPTION

[0003] The invention relates to the processing of medical diagnostic information, namely to forensic medicine, and can be applied to methods of identifying a person based on the biological activity of the heart by studying the parameters of electrocardiography.

[0004] The closest in technical essence is the METHOD FOR DETECTING AND IDENTIFYING PEOPLE BY CHARACTERISTIC SIGNALS [CN114545355 (A), published 05 / 27 / 2022], which includes receiving one or more monitored physiological parameters of a person from one or more sensors, wherein the monitored physiological parameters include cardiac parameters, processing signals from one or more sensors to identify a person based on an assessment of a biometric signature, processing including an assessment of the monitored parameters, including a cardiac parameter, determining one or more settings for a respiratory therapy device based on an assessment of the biometric signature.

[0005] The main problem of the prototype is that in this method, the identification of a person's personality is not performed by the parameters of the functioning of the cardiovascular system, registered from wireless communication channels received from smart devices, such as a smartphone, smart watch, smart Wi-Fi router, as well as the parameters of the state of wireless communication channels received from a cellular repeater and a base station channel, which significantly increases the accuracy of identification of a person by the parameters of wireless communication channels. In addition, modern methods of machine learning are not used to improve the accuracy of identification of a person's personality, which currently allow achieving high accuracy in identification tasks based on data of various physical origins.

[0006] The objective of the invention is to eliminate the shortcomings of the prototype.

[0007] The technical result of the invention is to increase the accuracy of human identification.

[0008] The specified technical result is achieved due to the fact that the method of contactless identification of a person based on the parameters of the functioning of the cardiovascular system is characterized by the fact that initially a training sample of the parameters of the functioning of the cardiovascular system is recorded, formed on the basis of the registration of electrophysiological signals, simultaneously received from electrocardiography and as a result of the registration of the parameters of the state of wireless communication channels, then a machine learning model is trained, within the framework of which the machine learning model is trained to identify a person by training to compare the results of patient diagnostics based on electrophysiological signals received from electrocardiography, the registered parameters of the state of wireless communication channels with a specific person,as a result, the machine learning model is trained to identify a person based on the results of the recorded electrocardiography parameters and the parameters of the state of wireless communication channels, then the trained machine learning model is used to identify the person.

[0009] In particular, to record the training sample of parameters of the functioning of the cardiovascular system, the RSSI and CSI parameters of the state of wireless communication channels are recorded.

[0010] In particular, if the trained machine learning model receives registered electrophysical signals and parameters of the state of wireless communication channels at the input, according to which the trained machine learning model does not produce a result of identifying the patient’s identity, then these registered parameters are fed to the machine learning model for retraining.

[0011] In particular, identification of a person’s identity using a trained machine learning model is performed based on the results of recorded electrocardiography parameters.

[0012] In particular, identification of a person’s identity using a trained machine learning model is performed based on the parameters of the state of wireless communication channels.

[0013] In particular, identification of a person’s identity using a trained machine learning model is performed based on the results of the recorded electrocardiography parameters and the parameters of the state of wireless communication channels simultaneously.

[0014] Implementation of the invention.

[0015] Typically, one or more biometric parameters such as fingerprints, voice prints, retinal scans, and facial features are used to identify a person. However, existing technical solutions do not use contactless electrocardiography parameters to identify a person.

[0016] The cardiac muscle is myogenic, i.e. it has the ability to contract rhythmically without any visible irritations under the influence of impulses arising in the muscle itself, which is a characteristic feature of the heart. Since impulses appear in muscle fibers, they speak of myogenic automatism. The myogenic regulation of the heart is based on the features of the structure and functioning of cardiomyocytes, the features of the connection of muscle cells with each other. That is, the cardiac muscle is capable of generating an action potential - an excitation wave moving along the membrane of a living cell in the form of a short-term change in the membrane potential in a small area of ​​the excitable cell (neuron or cardiomyocyte), as a result of which the outer surface of this area becomes negatively charged in relation to the inner surface of the membrane, while at rest it is positively charged.The action potential is the physiological basis of the nerve impulse, resulting in the formation of depolarization and repolarization signals within the heart muscle. The cardiac conduction system, consisting of a group of specialized cardiac cells (sinoatrial node; atrioventricular (atrioventricular) node; the bundle of His, extending from the atrioventricular node and dividing into two legs (right and left), each of which innervates the corresponding ventricle, and the left is also divided into anterior and posterior branches; Purkinje fibers: small nerve bundles extending from the branches of the legs of His), transmit the electrical signal through the heart. Based on the registration of statistical parameters of cardiac functioning, namely peak amplitudes and / or lengths of time intervals and / or angles of depolarization-repolarization vectors and / or lengths of depolarization-repolarization vectors for PQRST electrical signals associated with cardiac waves, it is possible to identify a person.Registered statistical parameters serve as biometric features for identifying a person. For this purpose, during registration, the raw electrocardiography (ECG) signal is processed, and the results in the form of parameters are serialized, i.e. the data structure is converted into a bit sequence and saved. During personal identification, the registered parameters are used to recognize (classify) subjects based on similar heart rate parameters extracted from the ECG signal of the person subject to verification or identification.

[0017] Experimental studies have shown that identification of a person based on the registration of cardiac muscle functioning parameters is much more accurate than identification based on facial parameters, and almost as accurate as identification based on a fingerprint.The method of contactless identification of a person based on the parameters of the functioning of the cardiovascular system is characterized by the fact that initially a training sample of the parameters of the functioning of the cardiovascular system is recorded, formed on the basis of recording electrophysiological signals, simultaneously received from electrocardiography performed using an electrocardiograph and obtained as a result of recording RSSI and CSI parameters of the state of wireless communication channels received from smart devices, such as a cellular (wireless) telephone, wireless multimedia accessories, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device, as well as RSSI parameters of the state of wireless communication channels received from a cellular repeater and a base station channel.It was experimentally established that the representativeness of the training sample was achieved with 1000 training standards formed during the registration of the parameters of the functioning of the cardiovascular system of 1000 people of different ages and genders, while it was established that the accuracy of personal identification does not depend on age, health status and gender.

[0018] Next, the machine learning model is directly trained, within the framework of which the machine learning model is trained to identify a person by training to compare the results of patient diagnostics based on electrophysiological signals obtained from electrocardiography, registered RSSI and CSI parameters of the state of wireless communication channels with a specific person, as a result of which the machine learning model is trained to identify a person both based on the results of the registered electrophysical signals obtained from electrocardiography, and based on RSSI and CSI parameters of the state of wireless communication channels.

[0019] Then, using the trained machine learning model, at the training stage of the machine learning model, the person's identity is identified, where the patient's identity is identified both by the registered electrophysical signals obtained from electrocardiography and by the RSSI and CSI parameters of the state of wireless communication channels received from smart devices, such as a cellular (wireless) telephone, wireless multimedia accessories, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device, as well as the registered RSSI parameters of the state of wireless communication channels received from a cellular repeater and a base station communication channel.If the trained machine learning model at the stage of training the machine learning model at the stage of identifying the person's identity receives registered electrophysical signals obtained from electrocardiography and parameters of the state of wireless communication channels at the input, according to which the trained machine learning model does not issue the result of identifying the person's identity, then these registered parameters are fed to retrain the machine learning model.

[0020] The technical result of the invention is an increase in the accuracy of human identification due to the fact that recording a training sample of the parameters of the functioning of the cardiovascular system, formed on the basis of recording electrophysiological signals obtained from electrocardiography and the corresponding parameters of wireless communication channels, allows to significantly improve the quality of training the machine learning model and, as a consequence, to increase the accuracy of human identification based on the parameters of the functioning of the cardiovascular system, while such an approach to training the machine learning model and subsequently identifying a person, carried out on the basis of registered parameters of both electrophysical signals and wireless communication channels, also allows to increase the accuracy of human identification.In addition, the accuracy of human identification during registration of wireless communication channel parameters is increased by retraining the machine learning model in cases where, during identification, parameters of electrophysical signals and wireless communication channels are registered that were not taken into account during training of the machine learning model. Examples of achieving the technical result:.

[0021] In example 1, a 58-year-old woman was identified, for whose identification, RSSI and CSI parameters of wireless communication channels received from smart devices were recorded, namely, a cell phone, smart watch, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device, as well as RSSI parameters of the state of wireless communication channels received from a cellular repeater and a base station communication channel, then using a pre-trained machine learning model, this woman was identified with an accuracy of 95%. Then, the identification of this woman was checked using the recorded electrophysical parameters obtained as a result of electrocardiography analysis, the identification accuracy was 85%.Further identification was performed on the basis of parallel recorded RSSI and CSI parameters of wireless communication channels received from smart devices, namely a cell phone, smart watch, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device, as well as RSSI parameters of the state of wireless communication channels received from a cellular repeater and a base station communication channel and recorded electrophysical parameters obtained as a result of electrocardiography analysis, the identification accuracy was 98%.

[0022] This example reflects the implementation of the claimed method. For this purpose, a group of 50 people aged 25 to 65 years was collected. For these people, training samples were recorded, initially formed on the basis of connecting the subjects to the electrocardiograph, and electrocardiography was recorded over four days, 5 electrocardiographs with a recording period of one hour each, as a result, 1000 electrocardiographs were collected, which were analyzed to identify electrophysical parameters (deviations of the main teeth: P, Q, R, S and T).

[0023] In parallel, when taking electrocardiographs, RSSI and CSI parameters of wireless communication channels were recorded, which were taken from smart devices, namely a cell phone, smart watch, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device and were recorded on a mobile phone: RSSI recorded parameters were the indicators of the power of incoming signals, interference and noise coming from base stations of cellular communication and repeaters, measured in dBm, and CSI recorded parameters were the parameters of the signal propagated from the transmitter to the receiver and reflecting scattering, fading and reduction of signal power, namely the parameters of instantaneous, statistical CSI.

[0024] Based on the obtained training sample, a machine learning model, namely a convolutional neural network, was trained. For this purpose, the training sample, consisting of data sets including parameters of electrophysical signals obtained as a result of analyzing electrocardiographs and recorded parameters of wireless communication channels, was divided into training - 50%, control - 30% and test 20%. As a result, the neural network was trained without retraining to correctly identify a person by the parameters of electrophysical signals obtained from the ECG, with a quality of 92%, by the parameters of wireless communication channels with a quality of 96%, by the parameters of electrophysical signals obtained from the ECG and wireless communication channels with a quality of 98%. For verification, a test sample was used, the data of which did not participate in training, but was used to assess the quality of training of the convolutional neural network.

Claims

AMENDED CLAUSE received by the International Bureau on December 4, 2024 (04.12.2024) 1. The method of contactless identification of a person based on the parameters of the cardiovascular system functioning is characterized by the fact that initially a training sample of the parameters of the cardiovascular system functioning is recorded, formed on the basis of recording electrophysiological signals, simultaneously received from electrocardiography and as a result of recording the parameters of the state of wireless communication channels, which were taken from smart devices, namely a cell phone, smart watch, wireless smart home devices, wireless tags, Wi-Fi router, wireless nearlink device and were recorded on a mobile phone: RSSI recorded parameters were the power indicators of incoming signals, interference and noise coming from cellular base stations and repeaters, measured in dBm, and CSI recorded parameters were the parameters of the signal propagated from the transmitter to the receiver and reflecting scattering, fading and reduction in signal power,namely, the parameters of instantaneous, statistical CSI, then the machine learning model is trained, within the framework of which the machine learning model is trained to identify a person by training to compare the results of patient diagnostics based on electrophysiological signals obtained from electrocardiography, registered parameters of the state of wireless communication channels with a specific person, as a result of which the machine learning model is trained to identify a person based on the results of the registered parameters of electrocardiography and based on the parameters of the state of wireless communication channels, during registration, the raw electrocardiography signal (ECG) is processed, and the results in the form of parameters are serialized, that is, the structure is translated, 13 AMENDED SHEET (ARTICLE 19) data into a bit sequence and save it, then use a trained machine learning model to identify a person.

2. The method according to item 1, characterized in that in order to record the training sample of parameters of the functioning of the cardiovascular system, the RSSI and CSI parameters of the state of wireless communication channels are registered.

3. The method according to item 1, characterized in that if the trained machine learning model receives registered electrophysical signals and parameters of the state of wireless communication channels at the input, according to which the trained machine learning model does not produce a result of identifying the patient’s identity, then these registered parameters are fed to retrain the machine learning model.

4. The method according to paragraph 1, characterized in that the identification of a person’s identity using a trained machine learning model is performed based on the results of the recorded electrocardiography parameters.

5. The method according to paragraph 1, characterized in that the identification of a person’s identity using a trained machine learning model is performed based on the parameters of the state of wireless communication channels.

6. The method according to paragraph 1, characterized in that the identification of a person’s identity using a trained machine learning model is performed based on the results of the recorded electrocardiography parameters and the parameters of the state of wireless communication channels simultaneously. 14 AMENDED SHEET (ARTICLE 19)