EP4373383A1

EP4373383A1 - System for assisting with provision of diagnostic information

Info

Publication number: EP4373383A1
Application number: EP22744446.0A
Authority: EP
Inventors: Georges De Pelsemaeker
Original assignee: Valeo Systemes Thermiques SAS
Current assignee: Valeo Systemes Thermiques SAS
Priority date: 2021-07-21
Filing date: 2022-07-05
Publication date: 2024-05-29
Also published as: FR3125405A1; WO2023001549A1; CN118055727A

Abstract

The present invention relates to a system for assisting with provision of diagnostic information, especially with a view to detecting a disease in a person, especially a contagious disease such as COVID-19, this detecting system comprising: - a device for acquiring examination data on the person, this acquiring device comprising at least two sensors for measuring a given physiological measurement, such as a radar and a thermal camera, to acquire these examination data, these sensors especially being arranged to work without contact with the person, - a data-processing device arranged to receive these data obtained by the sensors of the acquiring device, this processing device further being arranged to fuse data delivered at least by the two sensors of the acquiring device especially with a view to increasing the robustness of the examination data acquired on the person.

Description

HELP SYSTEM FOR PROVIDING INFORMATION ON

DIAGNOSTIC

[1] The present invention relates to an aid system for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19.

[2] The detection of potential COVID 19 patients is now carried out by taking the temperature, additional clinical examinations, then by a COVID 19 test which still has a relatively low level of confidence. A chest scan can with a good level of accuracy confirm the severity of the disease.

[3] There is a significant need to be able to quickly detect, in particular using mobile means, potential patients with a contagious disease, for example the disease linked to COVID 19.

[4] The subject of the invention is thus a system for detecting a disease on a person, in particular a contagious disease such as COVID 19, this detection system comprising:

[5] - a device for acquiring examination data on the person, this acquisition device comprising in particular at least one physiological measurement sensor such as a radar, and a thermal camera for acquiring this examination data,

[6] - a data processing device arranged to receive these examination data obtained by the acquisition device,

[7] - a display device arranged to display diagnostic information of the disease based on an analysis of said examination data, this diagnostic information being able to be representative of a level of probability that the person is suffering from the disease .

[8] According to one of the aspects of the invention, the data processing device comprises an artificial intelligence unit arranged to process the data examination data obtained by the acquisition device and providing said diagnostic information.

[9] According to one of the aspects of the invention, the data processing device and the display device are part of the same device, for example a computer, in particular a laptop computer.

[10] In this case, the display device is a laptop computer screen and the data processing device includes this computer's microprocessor.

[11] In a variant, the display device is arranged to be visible to the person examined, and in particular the display device, which notably comprises a screen, is remote from the data processing device, these display device and data processing being for example interconnected by a wireless link, for example by a 3G, 4G or 5G communication protocol, or by the Internet network, or Wifi for example.

[12] According to one of the aspects of the invention, the display device, and possibly also the data processing device, are arranged to be embedded in a motor vehicle.

[13] As a variant, the display device, and possibly also the data processing device, are arranged to be placed in a fixed manner, in a building or an outdoor courtyard for example.

[14] According to one of the aspects of the invention, the processing device is arranged to make the diagnosis automatically, without human intervention.

[15] Due to automatic processing, the invention allows rapid diagnosis and/or massive screening allowing a return to work or faster deconfinement.

[16] According to one of the aspects of the invention, the examination data acquisition device is arranged to allow non-contact and safety distance measurements of vital signals and thermal and visible imaging on the person to examine.

[17] The system according to the invention thus advantageously uses the fusion of measurements without contact and at a safety distance, of vital signals and of thermal and visible imaging. [18] According to one of the aspects of the invention, the artificial intelligence unit is arranged to use a diagnostic model based on artificial intelligence and fed by a reasonable number of clinical measurements.

[19] The invention, due to relatively light equipment, notably allows easy deployment of field hospitals in support of populations.

[20] The invention allows diagnostics to be carried out in a mobile manner. Its implementation, for example using a thermal camera, a physiological measurement sensor such as a radar and a portable personal computer, is easy. The invention allows its rapid implementation throughout a territory.

[21] According to one of the aspects of the invention, the acquisition device comprises a radar to acquire data relating to the vital signs of the person, a thermal camera for temperature measurements providing temperature data, and a camera operating in the visible spectrum for the characterization of the person tested providing data characterization of the person.

[22] According to one of the aspects of the invention, the data processing device is arranged to run a diagnostic algorithm based on a fusion of vital data, in particular a respiratory rate, an amplitude of respiration, a time of inspiration and expiration, heart rhythm and arrhythmia.

[23] According to one of the aspects of the invention, the algorithm uses temperatures measured on remarkable areas, located by image processing, or even the oxygen level linked to the person examined.

[24] Preferably, these remarkable temperature measurement zones are located inside the mouth, on the tip of the nose, on the cheekbones and the palm of the hand.

[25] According to one of the aspects of the invention, the diagnostic algorithm uses a characterization of the person, such as data on age, gender, clothing, height, body mass index , also called BMI.

[26] According to one of the aspects of the invention, the system is arranged to carry out the acquisition of examination data until the provision of the diagnostic information in a period of time comprised in particular between 30 and 120 seconds. [27] According to one of the aspects of the invention, the system is arranged to allow acquisition of examination data by taking measurements at a distance of 60 cm to 2 m between the acquisition device and the person. It is avoided that the person has to be in contact with the acquisition device.

[28] The invention thus allows rapid diagnosis without additional delay in getting to a doctor, for example. Diagnostic information can, if necessary, be automatically sent to a doctor and can be kept on a cloud-type data storage system.

[29] According to one of the aspects of the invention, the data processing device uses an algorithm for analyzing the acquired examination data and, if necessary, for sorting people with the aim of detecting cases of sick people, based on all the data collected and an artificial intelligence whose first level of learning is carried out on a sample in a hospital environment.

[30] This learning by artificial intelligence can be carried out by means of a set of measurements collected by the system but also by means of the medical follow-up of patients. This allows the improvement of the model over time.

[31] According to one of the aspects of the invention, the measurements taken by the acquisition device can be used to subsequently refine the diagnosis made by the artificial intelligence.

[32] Thanks to the invention, a diagnosis can be made without contact with the person, which limits the risks of contamination, which is particularly advantageous, for example, in the case of a pandemic such as that linked to COVID 19.

[33] According to one of the aspects of the invention, the measured examination data includes at least one of the following data: temperatures measured at different points of the body of the person to be examined, a respiratory or cardiac characteristic.

[34] According to one of the aspects of the invention, the acquisition device is arranged to acquire examination data comprising an outside temperature, a temperature measured on a cheekbone of the person, a temperature measured on the tip of the person's nose, and also if applicable, a maximum facial temperature and a temperature of a garment or a reference temperature-controlled surface.

[35] According to one of the aspects of the invention, the remarkable measurement points are located by an artificial intelligence by means of an object identification flowchart.

[36] According to one of the aspects of the invention, the temperature relative to a remarkable point is obtained by time average and by average of the temperatures of a surface defined by pixels taken from an infrared camera image near the remarkable point identified on the visible image by means of an object identification algorithm.

[37] According to one of the aspects of the invention, the identification of personal characterizations is done by means of Red Green Blue cameras (or RGB camera) or Far Infrared (FIR camera) in addition to an identity reading of the person, thanks to a classification system which can be learned on RGB (Red Green Blue) or Infrared images. The use of more parameters, including age, gender, height, body mass index, phenotype for example, serves to improve diagnostic models.

[38] According to one of the aspects of the invention, the diagnostic model, or diagnostic algorithm, which is fed with more data such as remarkable body temperatures, ambient temperature, a class of personal characteristics, a time of day, can be arranged to additionally use route data of the examined person to check whether he has come across a sick person or has passed through a region at risk.

[39] According to one of the aspects of the invention, the system is arranged to operate in the absence of a radar and using RGB cameras to estimate cardiac and respiratory parameters.

[40] According to one of the aspects of the invention, the temperature relative to a remarkable point can be obtained by time average and by average of the temperatures of a surface defined by pixels taken from a camera image near the remarkable point. The remarkable point is for example defined geometrically by means of an image area called Building box, which surrounds it, for example by means of the geometric mean of the sides of the image area. This image area is a surface delimited by a series of points which is constructed by an object identification algorithm.

[41] According to one of the aspects of the invention, the system does not have an RGB camera and/or does not use the surface whose temperature is controlled. In this case, the system uses the external temperature and a heat transfer model on the covered areas or even only the temperature differences between the remarkable points.

[42] According to one of the aspects of the invention, the system is arranged to use a fusion of non-contact measurements, in particular of vital signals and thermal and visible imaging.

[43] The diagnostic information is comprised of a class chosen from among three predetermined classes which are "Healthy person", "Person with a suspicion of disease", "Person with a high probability of disease".

[44] Diagnostic information may also include an assessment of disease severity.

[45] The invention also relates to a method for providing diagnostic information for the detection of a disease in a person, in particular a contagious disease such as COVID 19, this method comprising the following steps:

- acquiring examination data on the person, using an acquisition device comprising in particular at least one physiological measurement sensor such as a radar, and a thermal camera to acquire this examination data,

- receive this examination data obtained by the acquisition device,

- process this examination data to obtain diagnostic information for the disease based on an analysis of said examination data,

- displaying diagnostic information of the disease based on an analysis of said examination data, this diagnostic information being able to be representative of a level of probability that the person has the disease.

[46] The present invention can allow controls in the public space in general, in particular on people's circulation routes, at entrances and exits of buildings, at airport gates, in schools.

[47] This also allows medical follow-up of people, for example for sick people in home care.

[48] A further subject of the invention, independently or in combination with the foregoing, is an aid system for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system comprising:

- a device for acquiring examination data on the person, this acquisition device comprising in particular at least one sensor for physiological measurement such as a radar, and a thermal camera for acquiring this examination data, this device for acquisition comprising a plurality of sensors arranged to operate without contact with the person, at least one of the contactless sensors being arranged to acquire response data provided by the person,

- a data processing device arranged to receive these data obtained by several sensors of the acquisition device,

- a display device arranged to display diagnostic information of the disease based on an analysis of said data by the data processing device, this diagnostic information being able in particular to be representative of a level of probability that the person is suffering from the disease, this level of probability being able in particular to be a score.

[49] The invention allows, thanks to the plurality of sensors that operate without contact with the person to be examined, to have enough data to establish a more reliable diagnosis.

[50] According to one of the aspects of the invention, the data processing device comprises an artificial intelligence unit arranged to process the data examination data obtained by the acquisition device and providing said diagnostic information.

[51] According to one of the aspects of the invention, the data processing device and the display device are part of the same device, for example a computer, in particular a laptop computer.

[52] According to one of the aspects of the invention, the display device, and possibly also the data processing device, are arranged to be mounted on a motor vehicle.

[53] According to one of the aspects of the invention, the contactless sensor arranged to acquire response data provided by the person comprises a microphone arranged to receive responses provided orally by the person.

[54] According to one of the aspects of the invention, the contactless sensor can, as a variant of the microphone which works with voice recognition, include means for analyzing the person's face to, for example, recognize a YES or a NO, for example by nodding.

[55] As a further variant, the contactless sensor may include means allowing the person to point to an answer on a screen with a finger, without contact with the screen.

[56] According to one of the aspects of the invention, the system is arranged to present a questionnaire, for example by displaying the questions of the questionnaire on a display screen or for example using a sound device which pronounces the questions of the questionnaire, to the person to be examined, and the acquisition device is arranged to receive the answers provided by the person to this questionnaire, for example the microphone being used to record the person's answers.

[57] According to one of the aspects of the invention, the system comprises a data storage unit, for example forming part of a computer, arranged to store the questions of the questionnaire.

[58] According to one of the aspects of the invention, the data processing device is arranged to receive data from the acquisition device relating to the answers provided by the person in response to the questionnaire, and this device data processing is arranged to process these data in combination with data from other sensors of the acquisition device, in order to provide the diagnostic information of the disease.

[59] According to one of the aspects of the invention, in order to obtain the diagnostic information, the data processing device is arranged to run a diagnostic algorithm based on a fusion of the vital data of the person , including respiratory rate, amplitude of respiration, time of inspiration and expiration, cardiac rhythm and arrhythmia, with also questionnaire response data.

[60] According to one of the aspects of the invention, the system is arranged in such a way as to select the questions to be asked according to the person to be examined, in particular one or more data available on the person.

[61] According to one of the aspects of the invention, the questions to be asked are of a medical nature, useful for improving the diagnosis in combination with the other data.

[62] According to one aspect of the invention, the questions are chosen from the following list: Do you have chest pain? Do you have a dry cough? Do you have a sore throat? Do you have a runny nose? Do you have trouble breathing?

[63] According to one of the aspects of the invention, the system is arranged to perform, using one of the contactless sensors, a measurement whose result data is compared with an answer provided by the person in response to a question in the questionnaire.

[64] This makes it possible to make more reliable, or assign a level of reliability, to the measurement using the sensor.

[65] According to one of the aspects of the invention, the system is arranged to acquire data relating to a context, for example an ambient temperature.

[66] According to one of the aspects of the invention, the system is arranged to carry out a sorting based on a static model to give a score associated with the diagnosis.

[67] According to one of the aspects of the invention, the system is arranged to operate the following steps: receiving consent data from the person to be subjected to the examination, acquiring characterization data of the person such as gender, age, height, and data representative of the context, acquiring data in response to questions, in particular using a microphone, and measurements of vital signs, processing data to allow an evaluation of the level of reliability of the diagnostic information, to possibly ask additional questions if necessary to refine the diagnostic information, to carry out a sorting based on a static model to give a score associated with the diagnosis.

[68] According to one aspect of the invention, contactless sensors are arranged to allow measurement acquisition at a distance from the person to be examined, between 50 cm and 2 m.

[69] Another object of the invention is a terminal that includes the help system as mentioned above.

[70] This terminal can for example include a processing unit arranged to process data from the sensors and deliver diagnostic information.

[71] This processing unit can be a computer, with or without a screen on the terminal.

[72] This terminal notably includes the various sensors.

[73] This terminal can be installed at the entrance of a building, for example.

[74] Another subject of the invention is an aid method for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this method comprising the following steps:

- acquiring examination data on the person, this acquisition device comprising in particular at least one sensor for physiological measurement such as a radar, and a thermal camera for acquiring this examination data, this acquisition device comprising a plurality of sensors arranged to operate without contact with the person, at least one of the contactless sensors being arranged to acquire response data provided by the person, this sensor being in particular a microphone, - receive these data obtained by several sensors of the acquisition device,

- displaying diagnostic information for the disease based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person is suffering from the disease, this level of probability which may in particular be a score.

[75] Another subject of the invention is an aid system for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system including:

- a device for acquiring examination data on the person, this acquisition device comprising at least a first sensor, in particular a sensor for physiological measurement such as a radar or a thermal camera to acquire these examination data, in particular this acquisition device comprising a plurality of sensors arranged to operate without contact with the person, in particular at least one of the contactless sensors being arranged to acquire a response datum provided by the person,

- a data processing device arranged to receive these data obtained by the acquisition device,

- in particular a display device arranged to display diagnostic information of the disease based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person is affected of the disease, this level of probability being able in particular to be a score

- the acquisition device further comprising a correction sensor different from said first sensor and arranged to acquire at least one datum on the person, this datum being a correction datum different from data acquired by the first sensor, - the data processing device being arranged to use this correction datum to correct the examination data supplied by the first sensor.

[76] According to one of the aspects of the invention, the correction sensor is arranged to acquire data representative of a movement of the person.

[77] According to one of the aspects of the invention, this correction sensor comprises a camera, in particular of the 3D type, to acquire data representative of a movement of the person.

[78] This camera is in particular a camera operating on the principle of time of flight (in English: Time of Flight, TOF) which makes it possible to measure a scene in real time in 3 dimensions (3D). This type of camera is known.

[79] As a variant, this camera is an RGB camera operating with, for example, a body segmentation tool or a platform equipped with a strain gauge on which the person is installed, which person can be standing, lying or sitting.

[80] According to one of the aspects of the invention, the correction sensor is arranged to acquire data representative of a voice of the person.

[81] According to one of the aspects of the invention, this correction sensor arranged to acquire data representative of a voice of the person comprises a microphone (33), in particular a directional microphone or a camera.

[82] According to one of the aspects of the invention, the acquisition device comprises both a correction sensor arranged to acquire data representative of a movement of the person and a correction sensor arranged to acquire data representative of a person's voice.

[83] According to one of the aspects of the invention, the first sensor, in particular being a physiological measurement sensor, is a radar arranged in particular to acquire data representative of a breathing rate of the person, in particular by detecting the movement of the person's skin.

[84] According to one aspect of the invention, the radar is arranged to be placed in front of the person or behind the person's back. [85] According to one of the aspects of the invention, the data processing device is arranged to carry out the correction of the examination data according to the correction data using an artificial intelligence.

[86] According to one of the aspects of the invention, the data processing device is arranged to obtain a correction signal based on the correction data, this correction signal being evolving as a function of time.

[87] According to one of the aspects of the invention, the data processing device is arranged to provide the examination data, for example related to the respiratory rate, after correction using the correction signal.

[88] The invention makes it possible to take into account the noise related to the movements of the person and/or to his voice, to provide reliable and precise examination data.

[89] Indeed, the movements and words of the person are likely to create interference in the acquisition of data, in particular of physiological type such as the respiratory rate.

[90] Another object of the invention is a method for providing assistance with diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this method comprising the following steps:

- acquiring examination data on the person, using a device for acquiring examination data on the person, this acquisition device comprising at least a first sensor, in particular a sensor for physiological measurement such a radar or a thermal camera to acquire these examination data, in particular this acquisition device comprising a plurality of sensors arranged to operate without contact with the person, in particular at least one of the contactless sensors being arranged to acquire response data provided by the person,

- correcting this examination data using correction data, in particular acquired using a correction sensor.

[91] The invention has the advantage of being able to cross-reference data in order to obtain a more reliable and/or more robust result. [92] In the case of the evaluation of the heart rate by means of a radar, this evaluation being disturbed by the voice of the person or the movement of the person, the invention makes it possible to evaluate this disturbance thanks to a microphone or a camera that allows lip reading or even by means of a 4D camera or an inertial balance.

[93] The validity of the evaluation of this heart rate thus depends on the external noise, the posture of the person or the light intensity for example.

[94] According to one of the aspects of the invention, the method according to the invention includes the step of taking into account the weight of the evaluations of the parameters acquired to obtain the result.

[95] Thus when disturbances are of greater amplitude, the invention recommends giving more weight to the least disturbed parameters.

[96] According to one of the aspects of the invention, in the case of the evaluation of the heart rate by means of a radar, this evaluation being disturbed by the voice of the person or the movement of the person, the method presents the stage of evaluating this disturbance using a microphone or a camera that allows lip reading or even using a 4D camera or an inertial balance, to correct the examination data.

[97] Another subject of the invention is an aid system for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system including:

- a device for acquiring examination data on the person, this acquisition device comprising at least two sensors for measuring the same physiological measurement, such as a radar and a thermal camera for acquiring this examination data, these sensors being in particular arranged to operate without contact with the person,

- a data processing device (3) arranged to receive these data obtained by the sensors of the acquisition device, this processing device being further arranged to carry out a fusion of data coming from at least of the two sensors of the acquisition device, in particular with a view to increasing the robustness of the examination data acquired on the person,

- in particular a display device (30) arranged to display disease diagnostic information based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person has the disease, this level of probability being able in particular to be a score.

[98] According to one aspect of the invention, the two sensors are arranged at different locations.

[99] The two sensors can be arranged in the same box or in two different boxes.

[100] According to one of the aspects of the invention, the two sensors are arranged to pick up signals of different natures, for example signals of different wavelengths.

[101] According to one of the aspects of the invention, the two sensors comprising a radar and a thermal image camera are arranged to measure the respiratory rate or the heart rate of a person.

[102] Where “to measure respiratory rate” is stated above, it may refer to measurements of respiratory characteristics such as rate, amplitude, time of inspiration, expiration, apnea ...

[103] The invention makes it possible, thanks to the fusion of measurement data coming from two different sensors or more sensors, to have a more robust result on the diagnostic aid information.

[104] According to one of the aspects of the invention, one or more sensors are provided arranged to evaluate the noise of a signal and to attenuate this noise, in particular in addition to the evaluation of a level of reliability.

[105] According to one of the aspects of the invention, the processing unit is arranged to assign, for at least one of the sensors, a level of reliability, for example which can be 0 or 1, in one or more measurement windows, this reliability level and this window being used to perform the data fusion. [106] According to one of the aspects of the invention, the acquisition device comprises one or more complementary sensors arranged to determine the level of reliability of the two physiological measurement sensors.

[107] For example, in the case where the sensors are a radar and an infrared camera which make it possible to assess the breathing rate of the person, additional sensors such as a microphone or a device for measuring the movements of the person's body makes it possible to evaluate the reliability of the measurement made by the camera or the radar. In fact, when the person is moving or talking, the measurement made by the radar is either impossible or of poor quality.

[108] According to one of the aspects of the invention, this correction sensor comprises a camera, in particular of the 3D type, to acquire data representative of a movement of the person.

[109] For example, the infrared camera can be associated with an RGB camera to check whether the measurements by the infrared camera are possible or, when the person is moving or presents the head turned, for example, to conclude that the measurements by the infrared camera are impossible.

[110] According to one of the aspects of the invention, the processing unit is arranged to retain the measurement(s) of one or both sensors according to the reliability conditions of the measurement of the sensor(s).

[111] Thus the processing unit does not use the measurements made by the sensor, the reliability of which is not recognized as sufficient.

[112] For example, the processing unit does not use the measurements made by the sensor formed by a radar when the person is moving or turning his head. In this situation, the processing unit uses the other sensor which is an infrared imaging camera.

[113] According to one of the aspects of the invention, the processing unit is arranged to reconstitute a diagnostic information value on the basis of measurements obtained by the sensors and retained according to their quality and the reliability .

[114] For example, the curve of the signal representative of the diagnostic information can be formed by a signal measured by one of the sensors on a first time window and by a signal obtained by the other of the sensors over a second time window.

[115] This data fusion based on the reliability of sensor measurements and the consideration of time windows makes it possible to make the aid to obtaining diagnostic information robust.

[116] According to one of the aspects of the invention, the processing unit comprises an artificial intelligence arranged to carry out the data fusion.

[117] In particular, artificial intelligence is designed to assign a level of reliability to measurements made by the sensor(s).

[118] According to one of the aspects of the invention, the vital signals are evaluated by the artificial intelligence, and the uncertainty on the measurement evaluated by the artificial intelligence can be integrated statistically during the fusion of the data. .

[119] In particular, artificial intelligence is designed to accept or reject a measurement made by one or other of the sensors depending on the reliability of this measurement.

[120] Another subject of the invention is an aid method for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system including:

- providing a device for acquiring examination data on the person, this acquisition device comprising at least two sensors for measuring the same physiological measurement such as a radar and a thermal camera for acquiring this examination data, these sensors being in particular arranged to operate without contact with the person,

- receive these data obtained by the sensors of the acquisition device, this processing device being further arranged to carry out a fusion of data coming from at least the two sensors of the acquisition device, in particular with a view to increasing the robustness of the data examination acquired on the person

[121] The invention and its various applications will be better understood on reading the following description and examining the accompanying figures: [122] Figure 1 schematically illustrates a system according to a non-limiting embodiment of the invention.

[123] Figure 2 shows a block diagram illustrating the steps implemented in the system of Figure 1,

[124] Figure 3 illustrates a block diagram illustrating the steps implemented in a system according to another example of implementation of the invention,

[125] Figure 4 shows curves illustrating the effect of the corrections obtained thanks to the invention,

[126] Figure 5 shows curves illustrating the effect of merging data obtained using the invention.

[127] There is shown in Figures 1 and 2 an aid system 1 for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, according to the invention.

[128] This system 1 includes:

- a device for acquiring 7 examination data on the person,

- a data processing device 3 arranged to receive these examination data obtained by the acquisition device 7,

- a display device 30 arranged to display diagnostic information of the disease based on an analysis of said examination data, this diagnostic information being able to be representative of a level of probability that the person has the disease.

[129] This system 1 includes in particular:

- a sensor of the cardiac activity of at least one passenger, here the heart rate, this sensor being a camera operating in the near infrared,

- a respiratory activity sensor, in particular in frequency and/or respiratory amplitude, of at least one passenger, this sensor being a camera operating in the far infrared, or thermal camera,

- a radar arranged to measure the person's vital signs, - a sensor of the passenger's profile characteristics, in particular his gender, weight, height and age, this sensor being here a Red Green Blue camera, also called RGB camera in English,

- a card reader to read a person's identity card and obtain the person's personal data.

[130] These sensors and cameras, which are part of the acquisition device 7, are represented by the reference 2 in Figure 1. Some sensors 2 are for example arranged on a ceiling of the vehicle. One of the other cameras 2 is arranged in a side upright 6 of the vehicle V.

[131] The heart rate and respiration sensor may be in the seat back or in the center console at the level of the passenger's thigh, this being non-limiting.

[132] These sensors 2 are connected for exchange of information with a data processing device 3 placed on the vehicle V.

[133] The data processing device 3 includes an artificial intelligence unit arranged to process the examination data obtained by the acquisition device 7 and provide said diagnostic information.

[134] The data processing device 3 and the display device 30 are part of the same device, for example a computer, in particular a laptop computer.

[135] In this case, the display device 30 is a screen of the laptop computer and the data processing device includes this computer's microprocessor.

[136] The display device 30 is arranged to be visible to the person examined, and in particular the display device, which notably comprises a screen, is remote from the data processing device, these display and processing devices data being for example interconnected by a wireless link, for example by a 3G, 4G or 5G communication protocol, or by the Internet network, or Wifi for example.

[137] The display device 30, and also the data processing device 3, are here arranged to be carried on a motor vehicle. [138] Alternatively, the display device 30, and optionally also the data processing device 3, are arranged to be placed in a fixed manner, in a building or an outdoor courtyard for example.

[139] The processing device 3 is arranged to make the diagnosis automatically, without human intervention.

[140] The examination data acquisition device 7 is arranged to allow non-contact and safety distance measurements of vital signals and thermal and visible imaging on the person to be examined.

[141] The system according to the invention thus advantageously uses the fusion of measurements without contact and at a safe distance, vital signals and thermal and visible imaging.

[142] The artificial intelligence unit is arranged to use a diagnostic model based on artificial intelligence and fed by a reasonable number of clinical measures. [143] The invention allows diagnostics to be carried out in a mobile manner. Its implementation, for example using a thermal camera, a physiological measurement sensor such as a radar and a portable personal computer, is easy.

[144] The acquisition device 7 comprises a radar to acquire data relating to the vital signs of the person, a thermal camera for temperature measurements providing temperature data, and a camera operating in the visible spectrum for the characterization of the person being tested providing data characterizing the person.

[145] The data processing device 3 is arranged to run a diagnostic algorithm based on a fusion of vital data, in particular a breathing rhythm, a breathing amplitude, an inspiration and expiration time, a rhythm and cardiac arrhythmia, oximetry.

[146] The algorithm uses temperatures measured on remarkable areas, localized by image processing. [147] Preferably, these remarkable temperature measurement zones are located inside the mouth, on the tip of the nose, on the cheekbones and the palm of the hand.

[148] The diagnostic algorithm uses a characterization of the person, such as data on age, gender, dress, height, body mass index.

[149] The system is designed to carry out the acquisition of examination data until the provision of diagnostic information in a period of time comprised in particular between 30 and 120 seconds.

[150] The invention thus allows rapid diagnosis without additional delay in getting to a doctor, for example. The diagnostic information can, if necessary, be sent automatically to a doctor and can be kept on a data storage system 40 of Cloud type.

[151] The data processing device 3 uses an algorithm for analyzing the acquired examination data and, if necessary, sorting people with the aim of detecting cases of sick people, based on the set collected data and an artificial intelligence whose first level of learning is carried out on a sample in a hospital environment.

[152] The measurements made by the acquisition device can be used to later refine the diagnosis made by the artificial intelligence.

[153] Thanks to the invention, a diagnosis can be made without contact with the person, which limits the risks of contamination, which is particularly advantageous, for example, in the case of a pandemic such as that linked to COVID 19.

[154] According to one of the aspects of the invention, the measured examination data includes at least one of the following data: temperatures measured at different points of the body of the person to be examined, a respiratory or cardiac characteristic.

[155] The acquisition device 7 is arranged to acquire examination data comprising an outside temperature, a temperature measured on a cheekbone of the person, a temperature measured on the tip of the nose of the person, and also if applicable, a maximum face temperature and a reference temperature-controlled garment or surface temperature.

[156] The remarkable points of measurement are located by an artificial intelligence by means of an object identification flowchart.

[157] The identification of personal characterizations is done by means of Red Green Blue cameras (or RGB camera) or Far Infrared (FIR camera) in addition to a reading of the person's identity, thanks to a classification system whose learning can be performed on RGB (Red Green Blue) or Infrared images. The use of more parameters, including age, gender, height, body mass index, phenotype for example, serves to improve diagnostic models.

[158] The diagnostic model, or diagnostic algorithm, which is fed with more data such as remarkable body temperatures, ambient temperature, class of personal characteristics, time of day, can be arranged to additionally use journey data of the person examined to check whether they have crossed paths with a sick person or passed through a region at risk.

[159] The temperature relative to a remarkable point can be obtained by temporal average and by average of the temperatures of a surface defined by pixels resulting from a camera image near the remarkable point. The remarkable point is for example defined geometrically by means of an image area called Building box, which surrounds it. This image area is a surface delimited by a series of points which is constructed by an object identification algorithm.

[160] The diagnostic information is comprised of a class chosen from three predetermined classes which are "Healthy person", "Person with a suspicion of disease", "Person with a high probability of disease".

[161] The invention thus implements the following steps:

- acquiring examination data on the person using the acquisition device 7, which are steps 20 to 25 of Figure 2, - receiving these examination data obtained by the acquisition device,

- processing these examination data to obtain diagnostic information for the disease based on an analysis of said examination data, using the data processing device 3 (step 28),

- displaying diagnostic information for the disease based on an analysis of said examination data, this diagnostic information being able to be representative of a level of probability that the person has the disease (step 29).

[162] Steps 20 to 25 are as follows:

- the identification of personal characterizations is done by means of Red Green Blue cameras (or RGB camera), this is step 20,

- acquisition of the person's temperature in step 21 with an FIR type thermal camera, the details of this temperature acquisition being already described,

- acquisition of the respiratory rate step 22 using the FIR camera,

- acquisition of the heart rate in step 23 using the near infrared camera, or NIR camera,

- acquisition of vital signs using the radar, in step 24,

- acquisition of personal data using the card reader in step 25,

- potentially near-infrared (NIR) camera oximetry.

[163] This also allows medical monitoring of individuals.

[164] Examination data may, where appropriate, include pupil size and position.

[165] Diagnostic information is automatically sent to a cloud-based remote data storage system.

[166] Similarly, information useful for the diagnostic model, or diagnostic algorithm, can be received from the cloud-type remote system. [167] Potentially two cameras, NIR and FIR, are used. The FIR camera for temperatures and respiration characteristics and the NIR camera for oximetry and heart rate.

[168] The acquisition device is in particular arranged to acquire examination data comprising an outside temperature, a temperature measured on a cheekbone of the person, a temperature measured on the tip of the person's nose, and also, where appropriate , a maximum facial temperature and a temperature of a garment or of a reference temperature-controlled surface, and if applicable a respiratory volume, tremors, the level of oxygen in the blood.

[169] Figure 3 shows another example of implementation of the invention.

[170] In this example, the help system 1 includes, like the example described above:

- a device 7 for acquiring examination data on the person, this acquisition device comprising in particular at least one physiological measurement sensor such as a radar, and a thermal camera for acquiring this examination data, this device acquisition comprising a plurality of sensors 2 arranged to operate without contact with the person, at least one of the contactless sensors being arranged to acquire response data provided by the person,

- a data processing device 3 arranged to receive these data obtained by several sensors of the acquisition device,

- a display device 30 arranged to display diagnostic information for the disease based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person is affected of the disease, this level of probability possibly being in particular a score.

[171] One of the contactless sensors 2 arranged to acquire response data provided by the person comprises a microphone arranged to receive responses provided orally by the person. [172] The system 1 is arranged to present a questionnaire, for example by displaying the questions of the questionnaire on a display screen or for example using a sound device which pronounces the questions of the questionnaire, to the person to examine, and the acquisition device is arranged to receive the answers provided by the person to this questionnaire, here the microphone 33 being used to record the person's answers.

[173] System 1 comprises a data storage unit, for example part of a computer, arranged to store the questions of questionnaire 34.

[174] The data processing device 3 is arranged to receive data from the acquisition device relating to the answers provided by the person in response to the questionnaire, and this data processing device is arranged to process this data in combination with data coming from other sensors of the acquisition device, in order to provide the diagnostic information of the disease.

[175] With a view to obtaining the diagnostic information, the data processing device 3 is arranged to run a diagnostic algorithm based on a fusion of the person's vital data, in particular a respiratory rate, an amplitude of respiration , inspiration and expiration time, cardiac rhythm and arrhythmia, also with questionnaire response data.

[176] The system is set up in such a way as to select the questions to be asked depending on the person to be examined, in particular one or more data available on the person.

[177] The questions to be asked are of a medical nature, useful to improve the diagnosis in combination with the other data.

[178] The system 1 is arranged to carry out, using one of the sensors 2 without contact, a measurement whose result data are compared with an answer provided by the person in response to a question of the questionnaire.

[179] A description will now be given with reference to FIG. 3 of the various stages of implementation of the invention. [180] Step 50 corresponds to the beginning of the method for detecting a disease in a person.

[181] There follows a step 51 of data entry by the person to be examined.

[182] In this step 51, the person enters for example data concerning him, for example his identity, then possibly gives his agreement or consent to continue the diagnostic process.

[183] This step 51 can be performed by the person via a touch screen 35 coupled to a disinfection device 36 using UV. A microphone 33 can be used to record the person's responses to questions. A voice recognition device 38 is muted at the microphone 33 to input data picked up by the microphone 33 into the data processing unit 3.

[184] Context data 37, for example an ambient temperature, can be acquired by a sensor 2.

[185] In step 52, characteristics related to the person are determined using one or more sensors 2 via an object image recognition system 39, as described in the previous example.

[186] Additional questions are asked at the end of step 52 to clarify the person's clinical condition.

[187] For this step 53 of questions/answers, the questions are chosen for example from the following list: Do you have chest pain? Do you have a dry cough? Do you have a sore throat? Do you have a runny nose? Do you have trouble breathing?

[188] Answers can be given by the person via the microphone 33.

[189] In the following step 54, the sensors 2 proceed to the acquisition of data relating to the vital signs of the person, as described in the preceding example.

[190] Then in step 55, the artificial intelligence provides a preliminary score representative of the diagnosis obtained on the basis of the data entered until then. [191] At step 56, additional questions are asked of the person, and the answers obtained are used to refine the preliminary score obtained at step 55.

[192] A final score is given in step 57 by the artificial intelligence, which is recorded on the data storage space 40.

[193] At step 58, the final score is displayed for information to the person.

[194] We will now present another embodiment of the invention with reference to Figure 4.

[195] In this example, the aid system 1 for providing diagnostic information, slightly different from that described above, comprises:

- a device 7 for acquiring examination data on the person, this acquisition device comprising at least a first sensor, here a physiological measurement sensor which is a radar for acquiring this examination data,

- a data processing device 3 arranged to receive these data obtained by the acquisition device,

- a display device 30 arranged to display diagnostic information for the disease based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person is affected of the disease, this level of probability being able in particular to be a score

- the acquisition device 7 further comprising a correction sensor 33 different from said first sensor, namely the radar, and arranged to acquire at least one datum on the person, this datum being a different correction datum from data acquired by the first sensor,

- the data processing device being arranged to use this correction data to correct the examination data supplied by the first sensor.

[196] This correction sensor 33 is arranged to acquire data representative of a person's voice. [197] This correction sensor 33 arranged to acquire data representative of a person's voice comprises a microphone 33, in particular a directional microphone or a camera.

[198] Provision may be made for the acquisition device 7 to comprise both a correction sensor arranged to acquire data representative of a movement of the person and a correction sensor arranged to acquire data representative of a voice of the person.

[199] This correction sensor for acquiring data representative of a person's movement comprises a camera, in particular of the 3D type, for acquiring data representative of a person's movement.

[200] This camera is in particular a camera operating on the principle of time of flight (in English: Time of Flight, TOF) which makes it possible to measure a scene in real time in 3 dimensions (3D). This type of camera is known

[201] The radar is arranged in particular to acquire data representative of the person's breathing rate, in particular by detecting the movement of the person's skin.

[202] The radar is arranged to be placed in front of the person or behind the person.

[203] Of course, sensors other than radar can be used, for example a Near Infrared camera.

[204] The data processing device 3 is arranged to carry out the correction of the examination data according to the correction data using artificial intelligence.

[205] The data processing device 3 is arranged to obtain a correction signal based on the correction data, this correction signal being evolving as a function of time.

[206] The data processing device 3 is arranged to provide the examination data, for example related to the respiratory rate, after correction using the correction signal. [207] The invention makes it possible to take into account the noise linked to the movements of the person and/or to his voice, in order to provide reliable and precise examination data. [208] In fact, the movements and words of the person are likely to create interference in the acquisition of data, in particular of physiological type such as the respiratory rate.

[209] This can be seen for example on the curves in figure 4.

[210] Curve 101 illustrates the magnitude of the error, over time, in a breathing rate measurement on the person, this error being induced by movements of the person during the measurement and by the fact that she speaks during the measurement.

[211] This error amplitude on curve 101 is relatively large.

[212] The curve 102 shows the magnitude of the error, over time, on a measurement of the breathing rate on the person, when a correction is taken into account taking into account the movements of the person detected by the camera of movements.

[213] The amplitude on curve 102 is reduced compared to curve 101.

[214] Finally, curve 103 illustrates the amplitude of the error, over time, on a measurement of respiratory rate on the person, when the corrections linked to both the movements and the speech of the person are taken into account. no one.

[215] This curve 103 is close to zero thanks to the corrections made.

[216] The system of the invention thus makes it possible to implement the method comprising the following steps:

- correcting these examination data using correction data, in particular acquired using a correction sensor. [217] In connection with Figure 5, the system 1 according to an example of the invention may comprise

- the device 7 for acquiring examination data on the person, this acquisition device comprising at least two sensors for measuring the same physiological measurement such as a radar and a thermal camera for acquiring this examination data, these sensors being in particular arranged to operate without contact with the person, these sensors being already described above,

- the data processing device 3 arranged to receive these data obtained by the sensors of the acquisition device, this processing device being further arranged to perform a fusion of data coming from at least the two sensors of the acquisition device in particular by with a view to increasing the robustness of the examination data acquired on the person,

- the display device 30 arranged to display diagnostic information of the disease based on an analysis of said data by the data processing device, this diagnostic information possibly being representative of a level of probability that the person is affected of the disease, this level of probability possibly being in particular a score.

[218] The two sensors are placed at different locations.

[219] In the example described, the two sensors comprising a radar and a thermal image camera are arranged to measure a person's breathing rate.

[220] The processing unit 3 is arranged to assign, for at least one of the sensors, a level of reliability F, for example which can be 0 or 1, in several measurement windows T1, T2, T3 t T4 by example, this reliability level and these windows being used to perform the data fusion.

[221] These windows are visible on the graph in figure 5.

[222] The acquisition device 7 comprises one or more complementary sensors arranged to determine the level of reliability of the two physiological measurement sensors. [223] For example, in the case where the sensors are a radar and an infrared camera which make it possible to assess the breathing rate of the person, additional sensors such as a microphone or a device for measuring the movements of the person's body makes it possible to evaluate the reliability of the measurement made by the camera or the radar. In fact, when the person is moving or talking, the measurement made by the radar is either impossible or of poor quality.

[224] This correction sensor includes a camera, in particular of the 3D type, to acquire data representative of a movement of the person.

[225] For example, the infrared camera can be associated with an RGB camera to check whether the measurements by the infrared camera are possible or, when the person is moving or has his head turned, for example, to conclude that the measurements by the infrared camera are impossible.

[226] The processing unit 3 is arranged to retain the measurement(s) of one or both sensors depending on the reliability conditions of the measurement of the sensor(s).

[227] Thus the processing unit does not use the measurements made by the sensor, the reliability of which is not recognized as sufficient.

[228] For example, the processing unit does not use the measurements made by the sensor formed by a radar when the person is moving or turning his head. In this situation, the processing unit uses the other sensor which is an infrared imaging camera.

[229] The processing unit 3 is arranged to reconstitute a diagnostic information value on the basis of measurements obtained by the sensors and retained according to their quality and reliability.

[230] For example, the curve of the signal representative of the diagnostic information can be formed by a signal measured by one of the sensors over a first time window T1 and by a signal obtained by the other of the sensors over a second time window T3, as in the example in Figure 5.

[231] In windows T2 and T4, the measurements taken by the two sensors are of comparable satisfactory reliability and are merged. [232] In the graph of FIG. 5, the curve 110 represents the theoretical value of the information, for example the respiratory rate, which it is desired to obtain.

[233] Curve 111 represents a signal measured by the camera which is one of the sensors used to measure the breathing rate on the person.

[234] Curve 112 represents a signal measured by the radar which is the other of the sensors used to measure the breathing rate on the person.

[235] Curve 113 represents the fusion of the data represented by curves 111 and 112.

[236] As can be seen in Figure 5, in the T1 window, the reliability of the camera measurement is not sufficient and the measurement is not taken into account in the data fusion.

[237] In this window T 1 , only the measurements made by the radar, measurements represented by the curve 112, are taken into account.

[238] Conversely, in window T3, only the measurements taken by the camera are taken into account to obtain the result.

[239] In windows T2 and T4 for example, the camera and radar measurements, represented by curves 111 and 112, are sufficiently reliable to be all taken into account to obtain the resulting curve 113.

[240] This data fusion based on the reliability of sensor measurements and the consideration of time windows makes it possible to make the assistance in obtaining diagnostic information robust.

[241] The processing unit 3 includes an artificial intelligence designed to perform the data fusion.

[242] In particular, artificial intelligence is designed to assign a level of reliability to measurements made by the sensor(s).

[243] In particular, the artificial intelligence is set up to accept or reject a measurement made by one or other of the sensors depending on the reliability of this measurement.

Claims

[Claim 1] Aid system (1) for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system comprising:

- an acquisition device (7) for examination data on the person, this acquisition device comprising at least two sensors for measuring the same physiological measurement, such as a radar and a thermal camera for acquiring these examination data , these sensors being in particular arranged to operate without contact with the person,

- a data processing device (3) arranged to receive these data obtained by the sensors of the acquisition device, this processing device being further arranged to perform a fusion of data coming from at least the two sensors of the acquisition device in particular with a view to increasing the robustness of the examination data acquired on the person,

[Claim 2] System according to the preceding claim, according to which the two sensors are arranged in the same box.

[Claim 3] System according to one of the preceding claims, according to which the two sensors are arranged to pick up signals of different natures.

[Claim 4] System according to the preceding claim, according to which the two sensors comprising a radar and a thermal image camera are arranged to measure the respiratory rate or the heart rate of a person.

[Claim 5] System according to one of the preceding claims, according to which the processing unit is arranged to allocate, for at least one of the sensors, a level of reliability, for example being able to be 0 or 1, in one or more measurement windows (T 1 , T2, T3, T4), this level of reliability and this window being used to perform the data fusion.

[Claim 6] System according to one of the preceding claims, according to which the acquisition device (7) comprises one or more complementary sensors arranged to determine the level of reliability of the two physiological measurement sensors.

[Claim 7] System according to the preceding claim, according to which, in the case where the sensors are a radar and an infrared camera which make it possible to evaluate the respiratory rate or the heart rate of the person, additional sensors such as a microphone or a device for measuring the movements of the person's body making it possible to assess the reliability of the measurement made by the camera or the radar.

[Claim 8] System according to one of the preceding claims, according to which the processing unit is arranged to retain the measurement(s) of one or both sensors according to the conditions of reliability of the measurement of the sensor(s).

[Claim 9] System according to one of the preceding claims, according to which the processing unit is arranged to reconstitute a diagnostic information value on the basis of measurements obtained by the sensors and retained according to their quality and the reliability.

[Claim 10] System according to one of the preceding claims, according to which the processing unit comprises an artificial intelligence arranged to carry out the data fusion.

[Claim 11] Aid method for providing diagnostic information, in particular with a view to detecting a disease in a person, in particular a contagious disease such as COVID 19, this detection system comprising:

- providing an acquisition device (7) for examination data on the person, this acquisition device comprising at least two sensors for measuring the same physiological measurement, such as a radar and a thermal camera for acquiring these data from review, these sensors being in particular arranged to operate without contact with the person,

- receive these data obtained by the sensors of the acquisition device, this processing device being further arranged to carry out a fusion of data coming from at least the two sensors of the acquisition device, in particular with a view to increasing the robustness of the data examination acquired on the person.