Method and mobile terminal for quantitative measurement by chromatography of an analyte of interest

Abstract

The invention relates to a method (200) for quantitative measurement by chromatography of an analyte of interest, designed to be implemented by a mobile terminal. Such a method (200) comprises a step (210) of acquiring a stream of digital representations (20i) of a reactive strip delivered by the image sensor (22) and of selecting on the fly one of said digital representations (20i) as soon as it meets a predetermined set of relevance criteria. The method (200) further comprises a step (230) of producing a colouring intensity (I15i, I15', I15'') for each colouring profile (15i) expressed in a colour intensity distribution (100) obtained from the selected digital representation (20s'). The method (200) lastly comprises a step (240) of producing a measurement (M) of an analyte of interest based on the respective colouring intensities (I15', I15'') of two colouring profiles (15i).

Description

[0001] METHOD AND MOBILE TERMINAL FOR QUANTITATIVE MEASUREMENT BY CHROMATOGRAPHY OF AN ANALYTE OF INTEREST

[0002] technical field

[0003] The present invention relates to methods and systems for the quantitative measurement of analytes of interest by chromatography, said analytes being contained in liquid biological samples.

[0004] The term "analyte" refers to any substance, constituent, or biological, chemical, or biochemical entity whose presence in a liquid biological sample is to be quantified. Examples of such analytes include antibodies, hormones, proteins, peptides, enzymes, nucleic acids, and specific components (such as antigens or epitopes) of infectious agents like viruses or bacteria. Liquid biological samples in which the analyte may be in solution or suspension include any biological or bodily fluid from a human or animal subject, such as blood, plasma, serum, saliva, sweat, tears, cerebrospinal fluid, or urine.

[0005] Previous technique

[0006] The quantitative determination of an analyte of interest in a liquid biological sample by chromatography currently relies on the use of a test strip containing a detection or capture reagent specifically designed to identify the analyte of interest. This identification is manifested by a measurable color or fluorescence within a predefined detection zone, also called the "reactive zone," on the test strip. Parameters of this color, such as its intensity and / or extent, allow for the quantitative measurement of the presence of the analyte of interest in the liquid biological sample.

[0007] The identification of an analyte of interest by an inhomogeneous distribution of color intensities in the detection zone most often takes the form of identifying color profiles associated with the chromatographic technique used, such as a control line and a result line of the chromatographic test. Specifically, the quantification of the analyte being sought is most often directly related to the color intensity of the detected color profile(s). For example, the concentration of an analyte being sought is generally proportional to the color intensity of the corresponding color profile, hereinafter referred to as the "color profile intensity." Thus, a high color intensity is synonymous with a high concentration of molecules of interest.

[0008] In this regard, prior art exists for reading devices specifically designed to precisely determine the intensity of a color profile. Such a reading device is equipped with a receptacle that ensures a specific relative positioning (in terms of orientation, tilt, etc.) of a strip with respect to image capture means also integrated into the reading device, as described in US2021 / 0264604 A1. Controlling this relative positioning prevents any geometric distortion of the elements of interest during image capture. Furthermore, the size and position of each of these elements of interest within the captured image are also precisely determined.The image capture conditions of the test strip are also fully controlled and predefined by the reading device to prevent the presence of shadows or unwanted artifacts that could make chromatographic analysis of the strip impossible or compromise the validity of such an analysis. To this end, known reading devices also incorporate their own specific lighting systems to precisely and homogeneously illuminate the test strip being measured within the receptacle.

[0009] The combined arrangement of the reagent strip receptacle, capture devices, and lighting equipment during image capture can, however, depend on the type of reagent strip being analyzed. To accommodate this variety of reagent strips, it is known to use dedicated reading devices for specific types of reagent strips or, alternatively or in addition, to equip reading devices with a wide range of parameters. Calibrating these devices requires a specialist operator to adapt their functionalities to each type of reagent strip being analyzed. Thanks to this customization of the reading devices, the reagent strip analysis processes they employ are simplified, as they eliminate the need for image quality control (which is by definition optimal) and image correction.These images are, by their very nature, free of noise, interference, and distortion thanks to the design and specialization of the reading devices. However, such processes are unsuitable and inappropriate for use with images captured outside of such a controlled capture environment.

[0010] Such chromatographic analyses using specialized equipment are expensive due to the cost of designing, manufacturing, selling, and maintaining the reading devices, which require regular calibration, and remain the domain of specialist professionals. However, there is a clear need for widespread use of this type of analysis, whether at a patient's bedside, in the field, or at home.

[0011] Part of the journey toward facilitating biological testing in the field and making it accessible to everyone has been accomplished through the design of a microfluidic device for obtaining a plasma-reagent mixture, as disclosed in document FR3133922. Such a device primarily comprises:

[0012] - a blood collection module;

[0013] - a passive and non-forced plasma separation module from the collected blood, preventing any risk of hemolysis or the use of sophisticated third-party devices such as a centrifuge;

[0014] - a homogeneous mixing module for said plasma and a reagent;

[0015] - an output module for such a mixture to be deposited on an analysis strip.

[0016] Such a device clearly simplifies the process of collecting plasma and preparing a plasma-reagent mixture ready to be applied to test strips in the field, without requiring professionals or laboratory equipment. However, the need to transport these strips to a laboratory after application of the homogeneous mixture for precise analysis and interpretation of results by operators specializing in dedicated reading devices undermines this objective of democratization.

[0017] Document US2015 / 032006A1 confirms the need to be able to interpret a reagent strip directly at the patient's site without requiring dedicated equipment (hardware and software). This document thus envisions the use of a mobile phone or tablet. However, such a solution requires the use of a test-specific calibration reference or guide to determine whether a captured image is acceptable, or even improved. This particularly brief document provides little guidance on the actual feasibility of a reliable and accurate technical solution to meet this need. Document US2022 / 0084659 A1 also proposes using a mobile telecommunications terminal instead of a dedicated computer to perform certain operations in support of a remote analysis system.However, the document specifies the need for a particular kit to control the illumination of the test strip and the positioning of the optical sensor of the terminal relative to the test strip during image acquisition. This ensures that the terminal can provide a patient health profile by utilizing an optimal image of the test strip—that is, an image free of geometric distortion, shadows, or noise, and in which the size and position of the test strip's detection zone are known. Document US2021 / 0325299 A1, on the other hand, discloses a technique to assist a user in using a mobile phone to capture an image of a test strip revealing various water characteristics (pH, metal concentrations, etc.) by means of colored zones, and to compare these zones to a reference to deduce the respective values ​​of these characteristics.Such a reagent strip does not have a precise detection zone. The entire strip is read. Therefore, it is not necessary to precisely locate a detection zone encapsulated within a container containing a reading window, since such a reagent strip presents no risk of contamination by or for the operator. Consequently, there are no technical problems arising from the influence of this container on a captured image (shadows, reflections, impact on colors). Furthermore, the lesson learned from this document, which consists of comparing colored patches to a reference standard, does not produce a quantification of an analyte used to provide a diagnosis or inform a clinical decision.

[0018] None of the aforementioned documents discloses a technique for using a consumer-grade terminal to perform free capture of a test strip without specific lighting devices for said strip or positioning of the terminal's image sensor relative to it during image acquisition, while still allowing this same consumer-grade terminal to quantify an analyte accurately and reliably at a level comparable to the quantification that a specialized laboratory could perform on the same test strip. Prior art techniques produce approximate measurements or Boolean results of the "positive / negative" type without control over the capture conditions, or require the use of dedicated equipment to achieve greater analytical precision.

[0019] Therefore, there is a need for methods and systems for the quantitative measurement by chromatography of an analyte of interest, without the need to use laboratory material or human resources while maintaining high reliability of results.

[0020] Description of the invention

[0021] The invention proposes a method for quantitative measurement by chromatography of an analyte of interest, said method being designed to be implemented by a processing unit of a terminal comprising an image sensor and an output human-machine interface, said method comprising: - a step of producing a color intensity for each color profile expressed in a digital representation of a development zone of a reactive strip;

[0022] - a step of producing a measurement of an analyte of interest from the respective color intensities of two color profiles expressed in said digital representation of the development zone of said reagent strip.

[0023] To avoid the need for a dedicated, expensive terminal reserved for professional laboratory use, and to allow free image capture without having to resort to specific devices for illuminating the test strip and / or positioning it relative to the terminal's image sensor during the acquisition of the digital representation of the test strip's development area, such a process is arranged as follows:

[0024] - it comprises: o a step of acquiring a stream of digital representations of the reactive strip delivered by the image sensor and of selecting on the fly one of said digital representations as soon as it satisfies one or more relevance criteria; o a step of developing a working image from said selected digital representation, encoding a light intensity per pixel, said working image describing a distribution of color intensities of said development area of ​​said reactive strip in the form of a distribution of light intensities of the pixels forming said working image, said step of developing a working image comprising:

[0025] ■ a substep of extracting a region of interest within said selected digital representation including said reactive strip development area; ■ a substep of correcting geometric distortions in the selected digital representation with respect to an expected shape of a typical development area of ​​a reactive strip, said geometric distortions resulting from capture incidents during the acquisition step of a stream of digital representations;

[0026] - the step of producing a color intensity for each color profile expressed in a digital representation of a development area comprises: o a sub-step of modeling the distribution of color intensities expressed in the working image in the form of a color intensity curve describing the cumulative light intensity by each set of pixels aligned perpendicularly to a longitudinal axis of the development area; o a sub-step of identifying a color profile from said color intensity curve and determining a range of abscissas characterizing the beginning and end of said color profile; o a sub-step of calculating the color intensity of said identified color profile as being the area under the intensity curve on said determined range of abscissas;

[0027] - the step of producing a measurement of an analyte of interest: o consists of calculating a ratio between the respective color intensities of two color profiles; o causes an output of said measurement of an analyte of interest produced by the output human-machine interface of the terminal.

[0028] To analyze the entire development area of ​​a reagent strip, an instance of the substep of identifying a color profile from said color intensity curve and determining a range of abscissas characterizing the beginning and end of said color profile, as well as an instance of the substep of calculating the color intensity of said identified color profile, may be iterated from the lowest abscissa to the highest abscissa of said color intensity curve.

[0029] In order to reduce the negative impact induced by the presence of residual colour intensity resulting from less than optimal lighting conditions of the reactive strip during the acquisition step, the substep of calculating the colour intensity of said identified colour profile may include a step of subtracting from the area under the intensity curve on said determined abscissa range characterizing a colour profile, an area of ​​residual colour intensities.

[0030] In this case, the area of ​​residual color intensities of a color profile can consist of the area under a curve defined by an affine function linking the beginning and end of the color profile over said range of abscissas characterizing said color profile.

[0031] To maximize the relevance of the captured digital representations and automatically trigger the measurement process of the analyte of interest as soon as possible, the step of acquiring a stream of digital representations of the reactive strip delivered by the image sensor and selecting, on the fly, one of said digital representations may include:

[0032] - a substep of capturing a digital representation (20i);

[0033] - a substep of searching within said digital representation for a contour of an internal surface of a reactive strip comprising a development zone, and of selecting said digital representation: o if a surface contour describing a contour of a predefined geometric shape of the reactive strip surface has been identified, and o if a ratio between the area captured by such a surface contour and the total area of ​​said digital representation is greater than and at a predetermined threshold. To guide the terminal user to correctly position the terminal relative to the reactive strip during the step of acquiring a stream of digital representations, said terminal advantageously being a "consumer" mobile terminal, the step of acquiring a stream of digital representations of the reactive strip delivered by the image sensor and selecting, on the fly,one of said digital representations may include a sub-step of displaying it by the output human-machine interface of the terminal such that a suitable geometric shape describing a typical template of a "control" test strip is transparently superimposed on said digital representation in order to prompt the terminal user to position its image sensor relative to the test strip so that said suitable geometric shape covers the visible surface of the test strip on said digital representation displayed by said output human-machine interface.

[0034] To increase the relevance of on-the-fly selection of a digital representation, the relevance criterion or criteria that such a digital representation must satisfy to be selected on the fly may be chosen from a set of criteria including ambient brightness during capture, sharpness, contrast, color temperature of said captured digital representation and the content that the latter expresses.

[0035] To allow the adaptation and configuration of the production steps of a color intensity of a color profile as well as the production step of a measurement of an analysis, to different types of chromatographic test associated with a reactive strip, the step of developing a working image from said selected digital representation may include a substep of determining the type of chromatographic test associated with the reactive strip by implementing a machine learning algorithm.

[0036] According to a second object, the invention relates to a computer program product comprising program instructions, which, when written into a program memory of a terminal further comprising a processing unit, an output human-machine interface and a data memory, and interpreted or executed by said processing unit of the latter, cause the implementation of a method for quantitative measurement by chromatography of an analyte of interest according to the invention.

[0037] According to a second object, the invention relates to a mobile terminal comprising a program memory, a processing unit, an output human-machine interface and a data memory, the program memory of which contains the program instructions of such a computer program product.

[0038] Brief description of the drawings

[0039] The invention will be better understood and other features and advantages thereof will become apparent from the following description of particular embodiments of the invention, given by way of illustrative and non-limiting examples, and with reference to the accompanying drawings, among which:

[0040] - Figure 1 schematically illustrates an implementation of a quantitative measurement process by chromatography of an analyte of interest likely to be contained in a liquid biological sample received by a reactive strip according to various embodiments;

[0041] - Figure 2 schematically illustrates modules or functional elements of a mobile terminal enabling the implementation of the aforementioned process according to various embodiments;

[0042] - Figure 3 schematically presents functional steps of the aforementioned quantitative measurement process according to various embodiments;

[0043] - Figure 4 illustrates, through a graphical representation of a distribution of color intensities of a development area of ​​a reactive strip, the implementation of a step of color profile detection and production of a color intensity for each of the color profiles;

[0044] - Figure 5 illustrates, through a graphical representation of a distribution of color intensities of a development zone of a reactive strip, the implementation of a variant of such a step of detecting color profiles and producing a color intensity for each of the color profiles, in order to eliminate residual color intensities.

[0045] Detailed description

[0046] In order to simplify the description of the different embodiments which will be represented in the remainder of this description, the same references will be used on the different figures to designate the same elements or similar elements which are interchangeable with each other.

[0047] Figure 1 presents a reactive strip 1, in particular an immunochromatographic strip, comprising an outline 1 1 of an inner surface 12 of said reactive strip 1.

[0048] The peripheral contour 11 can be rectangular, rounded rectangular, ellipsoidal, generally rectangular with gripping notches, or, more commonly, elongated. In one embodiment, this contour 11 is the peripheral edge of the test strip 1, that is, the edge defining its shape.

[0049] In another embodiment, the contour 1 is a line that surrounds said inner surface 12 following the outer peripheral edge of the reactive strip 1. This line may mark the perimeter of the reactive strip 1 or be printed, embossed, woven or, more generally, marked in any other equivalent way on the surface of the reactive strip 1.

[0050] The inner surface 12 of the test strip 1 is designed to receive a liquid biological sample. This liquid biological sample can be deposited in a predefined application zone of the test strip 1 and migrate along it by capillary diffusion. For example, the test strip 1 is a lateral flow immunoassay (LFI). Alternatively, the test strip 1 can be immersed in the liquid biological sample. The test strip 1 is designed in a known manner to receive the liquid biological sample and to detect the presence of the analyte of interest by means of an inhomogeneous distribution of color intensities in the material, forming a detection zone 13.

[0051] The inner surface 12 of the test strip 1 thus includes at least one detection zone 13 intended to reveal the presence of an analyte of interest in a liquid biological sample received by the test strip 1. The detection zone 13 is designed to allow for the visual identification of an analyte of interest likely to be contained in the liquid biological sample received by the test strip 1. To this end, the test strip 1 includes a specific detection or capture reagent that enables the analyte contained in said liquid sample to be revealed in the detection zone 13. Alternatively or in addition, such a reagent may be present in the biological sample, through premixing, as proposed in document FR3133922.

[0052] A reagent strip may have multiple detection zones for the detection of multiple analytes of interest. Thus, depending on its structure, a reagent strip allows the determination of a single analyte (mono-analyte) or the determination of multiple analytes of interest.

[0053] In a non-limiting embodiment, the inner surface 12 of a test strip 1 includes a chromatographic test identifier 14. This identifier may be in graphic form, for example, as a barcode or QR code label (or "tag"). This identifier is, for example, associated with the analyte being sought and / or allows for tracking of the test strip 1 for improved results management.

[0054] In a non-limiting embodiment, the chromatographic test identifier 14 incorporates or is associated with position (or location) data within the inner surface 12 of the detection zone 13. The location of the detection zone 13 within the test strip 1 may, in fact, vary from one type of chromatographic test to another or from one brand of test strip 1 to another.

[0055] Unlike a dedicated reading device known in the prior art, the invention proposes using readily available equipment, for example, the advantageous form of a suitable smartphone 2, by implementing a mobile application designed for this purpose, to perform a quantitative measurement of an analyte of interest in a liquid biological sample received by the test strip 1. Such equipment 2 may, instead of or in addition to a mobile phone, be a tablet, a computer, or, more generally, a mobile device or terminal implementing a reading method according to the invention. It is sufficient that said equipment 2 has means for capturing the test strip 1 and / or for analyzing images of said strip 1 produced by a third-party device.

[0056] Referring to Figure 2, a terminal 2 comprises a processing unit 21 including one or more microprocessors or microcontrollers arranged to implement one or more computer programs. This processing unit 21 controls, via signals carried by a communication bus symbolized in Figure 2 by double arrows in single lines, electronic elements including a memory M. The latter comprises a data memory 24 and a program memory 23, said memories 23 and 24 possibly forming a single physical entity M.

[0057] The term "memory" refers to any computer memory, whether volatile or non-volatile. Non-volatile memory is computer memory whose technology retains its data even when no electrical power is supplied. It can contain data resulting from input, calculations, measurements, and / or program instructions. The main non-volatile memories currently available are electrically writable, such as EPROM (Erasable Programmable Read-Only Memory), or electrically writable and erasable, such as EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, SSD (Solid-State Drive), etc. Non-volatile memories are distinguished from so-called "volatile" memories, whose data is lost when power is removed.The main volatile memories currently available utilize RAM technologies ("Random Access Memory" according to Anglo-Saxon terminology or also called "vm" memory >>), DRAM (dynamic vm memory, requiring regular updating), SRAM (static vm memory requiring such updating during an electrical under-powering), DPRAM or VRAM (particularly suited to video), etc.

[0058] According to Figure 2, a terminal 2 also includes means of communication 26 with the outside world in the form of an input unit and an output unit. Thus, the terminal 2 can communicate with an application server 3 or a remote computer system 3. These means of communication 26 cooperate with the processing unit 21 and ensure wireless or wired proximity communication with any other remote electronic entity 3.

[0059] To operate, a terminal 2 generally includes an electrical power source 27, external or internal, such as one or more batteries. The processing unit 21 may also include means for controlling an input and / or output human-machine interface 25. An "output human-machine interface" is defined as any device, used alone or in combination, that outputs or delivers a graphical, haptic, auditory, or, more generally, human-perceptible representation. Such an output human-machine interface may consist of, but is not limited to, one or more screens, loudspeakers, or other suitable alternative means. An "input human-machine interface" is defined as a computer keyboard, a pointing device, a touchscreen, a microphone, or, more generally, any interface designed to translate a gesture or instruction issued by a human into control or parameter data.Advantageously, the input and output human-machine interfaces may constitute a single physical entity, for example when said terminal 2 consists of a smart mobile phone.

[0060] The operation of terminal 2 can be adapted by loading into its program memory 23 a computer program P containing instructions arranged to trigger, when executed by the processing unit 21, the implementation of a suitable process. This program can use any programming language and be in the form of source code, object code, or code intermediate between source and object code, such as in an interpreted, partially or fully compiled form, or in any other desirable form.

[0061] A terminal 2 includes an image sensor 22 enabling the acquisition of a stream of digital images 20 from the reactive strip 1. The image sensor 21 is, in particular, a camera capable of providing a stream of video images by series of matrix digital representations (better known by the English term "frames") at a predefined frequency, prior to a possible recording of all or part of these frames in the form of matrix digital images in data memory 24.

[0062] A "digital image" or "frame" is understood to be a set of pixels capable of being displayed on a display screen, such as the screen 25 of terminal 2. Each pixel can be defined by at least one parameter, such as a color and / or a light intensity level. A digital image is advantageously in the form of a computer image file, for example, in raster mode, vector mode, or pixel mode, the content of which can be processed by the processing unit 21 of terminal 2. Thus, the program instructions of a computer program P, when executed by said processing unit 21, allow the operation of terminal 2 to be adapted so that it processes digital representations 20 acquired by the image sensor 22 and determines a result of the chromatographic test, in the same way as a dedicated reading device according to the prior art.Advantageously, the processing unit 21 is configured to perform on-the-fly processing, that is, real-time or dynamic processing, of captured frames 20 without interrupting the image acquisition procedure of the reactive strip 1 and without requiring the storage in data memory 24 of a plurality of images resulting from the acquisition procedure. Indeed, a method 200, detailed later in connection with Figure 3, for the quantitative measurement by chromatography of an analyte of interest according to the invention may include on-the-fly selection of a frame if it satisfies one or more criteria necessary and sufficient to perform a relevant analyte measurement.The invention thus allows a non-professional user to use consumer-grade equipment, such as their smartphone 2, to capture a stream of images from a test strip 1 under various lighting conditions and relative positions of their device 2 with respect to the test strip 1. As soon as a satisfactory frame is selected by their device 2, the analyte measurement is initiated. Conversely, the device 2 prompts the user, via its output human-machine interface 25, to continue acquiring the image stream until such a frame is automatically selected by their device 2, so that said user U can orient and / or position their device 2 with respect to the test strip to modify the capture conditions and / or ambient lighting.The display screen 25 of terminal 2 thus allows visualization of the stream of digital images 20 being acquired by the image sensor 21 and / or content generated by the processing unit 21 to guide the user U in his image acquisition procedure.

[0063] Referring now to Figure 3, a method 200 for the quantitative measurement by chromatography of an analyte of interest implemented by the processing unit 21 of a mobile terminal 2 such as that illustrated in Figure 2, comprises, like a method implemented by a dedicated reading device according to the prior art: a step 210 for image acquisition of a reagent strip 1; a step 230 for producing a color intensity 115i for each identified color profile 15i; - a step 240 for producing a measurement M of an analyte of interest from the respective color intensities 115', 115” of two identified color profiles 15' and 15”.

[0064] Such a method 200 according to the invention differs, however, from those known in the prior art in several respects. First, instead of an image acquisition step 210 consisting of capturing an image of the reactive strip under optimal capture conditions, the acquisition step 210 of a method 200 according to the invention consists of acquiring a stream of digital images 20 of such a reactive strip 1 and selecting a digital representation 20s of said reactive strip from said stream 20. Such a selection can advantageously be performed on the fly by selecting a frame if the latter satisfies one or more predetermined relevance criteria.

[0065] Furthermore, given the random conditions of captures, due to a free image acquisition procedure via a consumer-grade material, a measurement method 200 according to the invention includes a step 220 of developing a working image 20s' from said digital representation selected 20s in step 210, working image 20s' which describes a development area 13 of said reactive strip 1 expressing a distribution of colour intensities.

[0066] Given the uncontrolled and suboptimal capture conditions, prior to the production 230 of a color intensity 115i for each color profile 15i expressed in said working image 20s', said production 230 consists of identifying such color profiles 15i by exploiting the color intensity distribution of said working image 20s'. The production of a color intensity 115i for an identified color profile 15i is also suitable.

[0067] Finally, step 140 of producing a measurement M of an analyte of interest from the respective unit color intensities 115', 115" of two identified color profiles 15' and 15" can be similar to those known from the prior art. However, thanks to the invention, such a measurement M is provided in real time to the user via the output human-machine interface 25 of the mobile terminal 2, which was also used for the free acquisition of images of the reactive strip 1.

[0068] A method 200 according to the invention therefore includes a step 210 of acquiring a stream of digital images 20 of a reactive strip 1 by the image sensor 22 of the mobile terminal 2. Such a step 210 includes a first substep of capturing 201 of a frame whose display 202 by the human-machine interface 25 is triggered, preferably, concurrently with the capture 201 by the processing unit 21 of the terminal 2. To assist or guide the user U in acquiring images of a reactive strip 1, a suitable geometric shape to describe the typical template of a "control" strip can be superimposed transparently on the frames displayed by the output human-machine interface 25 during the capture of the frames to facilitate the gestures of the user U.According to this advantageous embodiment, the user U of the terminal 2 is guided and encouraged to correctly position his mobile terminal 2 with respect to the reactive strip 1 so that said appropriate geometric shape covers the visible surface of the reactive strip 1. A positioning will be judged correct when the user naturally orients the image sensor 22 of his terminal 2 so that the reference template covers, or even aligns with, a visible surface of the reactive strip 1 on the displayed frame.

[0069] Step 210 includes, in real time and on the fly, a substep 203 for selecting a frame 20i from the stream of digital representations 20 delivered by the image sensor 22. To be selected, such a current digital representation, or frame 20i, must satisfy one or more predefined relevance criteria for subsequent implementation of an analyte measurement step 240. Such relevance criteria relate, but are not limited to, and by way of example, to capture conditions such as ambient brightness, sharpness, contrast, and color temperature of the captured image, as well as to the content it expresses. One such content criterion relates, for example, to the geometry, shape, and dimensions of a captured object and / or the presence of a graphic label also present on such an object detected in the frame.Thus, if a frame 20i is deemed too blurry, too dark, or conversely overexposed, or if its content exhibits excessive non-linear geometric distortion, or if the frame 20i does not describe any shape resembling a reactive strip (a situation illustrated by link 203-3 in Figure 3), the frame 20i is ignored and not selected. Step 203 examines the next frame within the stream 20. As soon as a frame 20i meets these relevance criteria (a situation illustrated by link 203-y), it is selected, and acquisition step 210 ceases in favor of step 220, which produces a working image from which an analyte measurement can be performed.

[0070] To detect the presence of a digital representation of a reactive strip 1 within a frame 20i, substep 203 may consist of a search for a contour of an interior surface 12 of said reactive strip 1 comprising a revealing area 13.

[0071] In one embodiment, said substep 203 can be arranged to search for a contour 1 of a predefined geometric shape of a reactive strip surface, for example, a substantially rectangular contour. In an advantageous embodiment, a Hough transformation can be applied to said detected contour 1 to create straight contour lines. Such substep 203 can then consist of calculating a ratio between the area captured or delimited by such a detected surface contour and the total area of ​​said acquired digital representation 20i. A ratio that is too low, for example less than 30%, or too high, for example greater than 90%, will be such as to reject the frame 20i, automatically considered to have been captured at a distance separating the reactive strip 1 from the capture means 22 of the terminal 2 that is unsuitable for producing a relevant digital representation of the reactive strip 1.Such a check advantageously prevents image acquisition 201 from being performed at a shooting distance that is too great or too short from the reactive strip 1, resulting in images of insufficient quality (noise, blur, excessive pixelation, etc.). To detect a reactive strip 1 contour 1, substep 203 may consist of querying a database (stored, for example, in the data memory 24 of terminal 2) describing the properties and / or templates of a plurality of commercially available reactive strips. These properties may include the two-dimensional shape of a reactive strip viewed from above, said shape being rectangular, with rounded corners, ellipsoidal, having one or more notches or gripping shoulders, concave or convex portions, etc.Such properties can, alternatively or in addition, describe a ratio between two of the dimensions of a reactive strip or, more generally, any geometric property or distinctive feature (a symmetry, a color and / or a marking for example) characteristic of a reactive strip among its peers.

[0072] Alternatively or in combination, such a substep 203 may exploit a pattern recognition algorithm, for example, a machine learning technique, an artificial neural network, alone or in combination with a Hough transform, to detect such a contour 11 of a reactive strip representation within a 20i frame. Such a substep 203 may further consist of implementing a machine learning algorithm to compare the acquired digital 20i frame to reference images, etc. The invention is not limited to the aforementioned examples of techniques that can be used to select a relevant 20s frame.

[0073] Thus, substep 203 advantageously consists of an automatic selection of a 20s frame retained for the continuation of the implementation of process 200 or, in other words, an automatic pre-sorting of the 20i frames of the image acquisition stream 20.

[0074] According to an advantageous embodiment of a method 200 according to the invention, when no frame taken from the acquisition stream 20 can be selected because all fail the relevance test carried out by sub-step 203, the latter can trigger the implementation of a sub-step 204 arranged to cause the output, via the output human-machine interface 25 of the terminal 2, of a help message or more generally, an invitation to the user suggesting to the latter a repositioning of the terminal 2 with respect to the reactive strip 1, a modification of the ambient lighting and / or a reconfiguration of the image sensor 22.

[0075] As mentioned previously, a method 200 according to the invention includes a step of developing a working image 20s' from a selected digital representation (frame) 20s describing a development area 13 of said reactive strip 1.

[0076] The purpose of such a working image is to express as accurately as possible the color intensity distribution 100 of a development area 13 of a reactive strip. This step 220 is crucial because the image acquisition procedure 210 is free and open to any user of "consumer" equipment 2, such as a smartphone. Unlike state-of-the-art techniques using dedicated reading devices, the capture conditions cannot be optimal, although the selection 203 may have rejected any frame too far removed in terms of relevance for the subsequent implementation of the process 200.

[0077] Thus, the use of a "general public" mobile terminal 2 inevitably leads to the presence of imperfections in the selected frame 20s, such as shadows, artifacts, parallax effects resulting from capture incidents during the acquisition step 210. The step 220 therefore consists of correcting the selected image 20s to produce a suitable working image 20s' to conduct the subsequent analyte measurement step 240.

[0078] At the end of selection step 210, the selected image or frame 20s contains a graphic and digital representation of a strip 1. Step 220 then includes a substep 221 for extracting a region of interest within the image 20s, comprising the development area 13 of the reactive strip 1. A working image 20s' can be obtained after recentering said region of interest (mainly including the transcription of the development area 13) and cropping to remove the area outside said region of interest. Such a substep 221 can also consist of creating a grayscale working image 20s' encoding a light intensity per pixel.Advantageously, prior to such a grayscale conversion, the area of ​​interest can be deconvolved from the selected digital representation 20s, so that a dominant color best expresses the reveal area 13, particularly the color profiles 15, in relation to the rest of the image. Such a substep 221 can thus rely on the singular value decomposition technique, also known by the abbreviation SVD. This decomposition aims to better separate the signal of the color profiles 15 (for example, in blue or red) from the rest of the reveal area (generally white).

[0079] Such a substep 221 may further consist of the application of a known technique aimed at implementing elementary operations (dilations then erosions) to reduce noise in the working image 20s' obtained.

[0080] Step 220 may further include a substep 222 for correcting any distortion in the selected frame 20s or in the image 20s' resulting from substep 221. Such distortion can manifest as a geometric deformation of the inner surface 12 of the captured digital representation of the strip 1, whose expected and predefined shape is generally rectangular, whereas this captured inner surface 12 has concave sides and / or elongated vertices or is still trapezoidal in shape. Several factors can cause distortion in an image 20s or 20s'. For example, approximate and free spatial positioning of the mobile terminal 2 relative to the reactive strip 1 can induce parallax errors.Similarly, the optics of the capture means 22 can also cause distortion (cushion or fish-eye effect) in the digital representation of said strip 1 from the capture step 210. Substep 222 can therefore consist of correcting the digital representation of the development zone 13, which is somewhat distorted. Such a correction can be achieved by implementing a trained machine learning algorithm equipped with a database containing distortion-free images of reactive strips. This algorithm thus makes it possible to determine and correct any detected distortion in order to restore a digital representation that most closely matches reality.

[0081] Thus, said substeps 221 and 222 are more generally arranged to detect and correct any geometric and / or colorimetric aberration presented by the image from the selected frame 20s and produce a working image 20s' describing in a relevant way a development zone 13 expressing one or more color profiles 15i, suitable for use by the rest of process 200.

[0082] A method 200 for the quantitative measurement by chromatography of an analyte of interest implemented by the processing unit 21 of a mobile terminal 2 such as that illustrated in Figure 2, according to the invention, comprises a step 230 of producing a color intensity 115i, 115', 115” for each color profile 15i, 15', 15” expressed in a digital representation 20s' of said test strip 1, in this case within the working image 20s' produced previously in step 220. The implementation of such a step 230 can be iterated as many times as there are development zones 13 within said image 20s' or as many working images 20s' associated with each development zone 13 of a test strip, if the latter comprises several.

[0083] Advantageously, but not limitingly, a working image 20s' describes a development area 13 as a digital matrix representation, oriented so that any longitudinal axis of said development area (generally in the form of a band) is horizontal. Thus, one or more color profiles 15 describe lines or bars of more or less homogeneous color oriented perpendicular to a reading axis AL of the development area 13.By convention, to illustrate the operation of a process 200 according to the invention, we will consider that the image 20s' produced in step 220 is such that the first color profile 15' revealed is a color profile called a "control profile," while a second color profile 15" reveals the presence of an analyte when a development zone 13 is scanned from left to right along a reading axis AL, generally longitudinal, of said development zone. The number of color profiles 15 in a development zone 13 can be predefined by the user as a predetermined parameter or be determined by the type T1 of chromatographic test associated with the reagent strip 1 if said type is automatically identified.As described previously, this test strip 1 can be determined by any characteristic specific to it (shape, dimensions, chromatographic test identifier 14, color, QR code, dimensions or position / orientation of the development zone 13 in the inner surface 12, reading axis AL). Its type Tl can thus be determined, in an optional substep 223, by means of a machine learning algorithm, said type Tl serving as a configuration parameter for steps 230 and 240 of process 200.

[0084] To ultimately produce a measurement related to an analyte of interest, a process 200 according to the invention includes a step 230 for generating a color intensity for each color profile identified in the working image 20s' representing the development zone 13 of the reagent strip. Such a step 230 includes a substep 231 for modeling the color distribution expressed by said development zone 13 within a working image 20s' in the form of a color intensity curve 100. More specifically, such a substep 231 consists of translating said revealing area 13 into the form of a histogram of light intensities for which the amplitude of each histogram bar translates the sum of the light intensities of the pixels of the same column of the matrix image 20s' when one traverses such an image 20s' from left to right along a reading axis AL of the revealing area 13.Such an intensity histogram can also be expressed as a linear curve 100 whose ordinates describe cumulative color intensities for a column of pixels in the image 20s' and whose abscissas correspond to these same columns of pixels considered from left to right along such an axis AL of the development zone 13. We will subsequently refer to this as an "intensity curve" 100 to characterize the result of such a substep 23s of modeling the digital representation of a development zone 13 of a reactive strip 1, in this case the working image 20s'. Depending on the presence or absence of analytes, such an intensity curve 100 exhibits a plurality of slopes or peaks characterizing respectively one or more color profiles 15.

[0085] Thus, Figure 4 presents an example of such an intensity curve 100 produced in a substep 231 of a process 200 according to the invention. Such a curve 100 describes a first profile 15' (control profile), strongly marked by the presence of one or two partially overlapping peaks, high light intensities (abscissas between 150 and 270), and a second color profile (measurement profile) 15” (abscissas between 450 and 550) of lower amplitude. Such a curve 100 illustrates the non-homogeneity of the color profiles 15 present in a development zone 13 of a reagent strip 1. Classically, a color profile is represented overall as a bell-shaped or Gaussian distribution whose mean is the center of said color profile and reflects the average intensity of said color profile, and whose standard deviation describes the width of said bell-shaped distribution.

[0086] The step 230 of producing a color intensity for each color profile 15 expressed in the working image 20s' expressing the development area 13 of the reactive strip 1, thus iteratively comprises, that is to say as long as (situation illustrated by link 234-n on figure 3) the working image 20s' reveals a color profile 15i, by traversing said image 20s' from left (low abscissas) to right (high abscissas), two instances of substeps 232 and 233, respectively of identification of said color profile 15i and of calculation of its color intensity 15i. Thus, according to the example illustrated by figures 3 and 4, two colour profiles 15' and 15” were identified from the intensity curve 100 produced in sub-step 231. Each instance of a sub-step 232 is followed by an instance of a subsequent sub-step 233 of calculating the intensity 115i of each colour profile 15i identified in 232.Such a calculation 233 of the intensity of a color profile 15 can consist, in a simplified way, of calculating the area under the intensity curve 100 for the range of abscissas characterizing said color profile 15. Thus, in Figure 4, the range of abscissas 15'ar characterizes a first color profile 15' and the range 15”ar characterizes a second color profile 15”. The number of iterations of substeps 232 and 233 can therefore vary from one reagent strip to another, that is to say, according to the number of color profiles 15i that can express a development zone 13 of a reagent strip 1.

[0087] According to a first embodiment, substep 232 consists of fitting a Gaussian distribution (the 15'G Gaussian for the 15' color profile in Figure 4) or a mixture of Gaussians (the set 15'Gm of the 15'G1 and 15'G2 Gaussians for the 15' color profile in Figure 4) to the curve 100 for each color profile. Substep 233 then consists of calculating the integral of said fitted Gaussians. We can observe in the example illustrated by Figure 4 that a 15' color profile can exhibit a single maximum, in this case the 15'M maximum for the 15' profile, or several maxima, in this case two maxima, 15'Ma and 15'Mb, for the 15' profile. This peculiarity stems from the inhomogeneity of coloration within a coloration profile.Substep 232 considers the set of 15'GM Gaussians as defining a single profile because the respective means of the two Gaussians 15'G1 and 15'G2 of said set or mixture of Gaussians are not separated by a sufficient step (i.e., by a sufficiently wide range of abscissas) to characterize two distinct color profiles. The fact that said two Gaussians partially overlap corroborates this hypothesis.

[0088] Alternatively, a substep 232 aimed at identifying a color profile 15i may consist of detecting a range of abscissas that characterizes it, that is, detecting a "start" 15'b, 15"b and an "end" 15'e, 15"e, of a distribution substantially bell-shaped or a mixture of bells. For this purpose, such a substep 232 may implement a calculation of the gradient of the intensity curve 100. The calculated gradient advantageously allows the identification of notable points such as local or global extrema (minimums or maxima), the average slope of the intensity curve 100, or the slope of said intensity curve 100 between two values ​​of interest. In particular with the calculated gradient, the maximum values ​​15'Ma, 15'Mb and 15”M allow us to identify the core or average of each 15', 15” color profile and / or the number of these 15 color profiles revealed.The minimum values ​​around these means allow us to obtain an estimate of the abscissa ranges 15'ar and 15”ar representing respectively the said 15' and 15” colour profiles on the example in Figure 4. Substep 233 then consists of calculating for each detected 15' and 15” colour profile the area under the intensity curve 100 for each determined 15'ar and 15”ar abscissa range.

[0089] In connection with Figure 4, the invention provides a particularly advantageous embodiment for determining the beginning of a color profile, especially when it has a plurality of maxima, such as the 15' color profile, and thus the lower bound of the abscissa range that characterizes it. According to the example illustrated in Figure 4, the first 15' color profile can be modeled from a mixture 15'GM of a first Gaussian 15'G1 and a second Gaussian 15'G2.

[0090] Thus, substep 232 may consist of:

[0091] - to determine the lower bound of the 15'ar range characterizing the 15' color profile, the selection of an abscissa closest to that of a maximum between the mean of a first Gaussian distribution 15'G1 of the mixture of Gaussian distributions 15'GM and the lower bound of an interval defined by a predefined number of standard deviations (for example, three) of a second Gaussian distribution 15'G2 of said mixture of Gaussian distributions 15'GM centered around the mean of this second Gaussian distribution 15'G2, the mean of the first Gaussian distribution 15'G1 being the smallest mean among the means of the Gaussian distributions of said mixture 15'GM, the mean of the second Gaussian distribution 15'G2 being the second smallest mean among the means of the Gaussian distributions of said mixture 15'GM;

[0092] - to determine the upper bound of the 15'ar range characterizing the 15' color profile, the selection of a second abscissa close to the minimum between the mean of the Gaussian distribution 15'G2 of said mixture 15'GM and the upper bound of an interval defined by a predefined number of standard deviations (for example, three) of the first Gaussian distribution 15'G1 of said mixture 15'GM centered around the mean of this first Gaussian distribution 15'G1, the mean of the Gaussian distribution 15'G2 being greater than the mean of the Gaussian distribution 15'G1, the mean of the Gaussian distribution 15'G2 being the second smallest mean among the means of the Gaussian distributions of said mixture 15'GM having a height less than a first predefined threshold value.

[0093] More generally, to determine the upper bound of such a range of abscissas 15'ra characterizing a coloring profile modelable by a mixture 15'GM of Gaussian distributions containing a number n greater than or equal to two, such a substep 232 may consist of selecting an abscissa close to the minimum between the mean of the Gaussian distribution 15'Gn of said mixture 15'GM and the upper bound of an interval defined by a predefined number of standard deviations (for example, three) of the Gaussian distribution 15'Gn-1 of said mixture 15'GM centered around the mean of this Gaussian distribution 15'Gn-1, the mean of the Gaussian distribution 15'Gn being greater than the mean of the Gaussian distribution 15'Gn-1, the mean of the Gaussian distribution 15'Gn being the second smallest mean among the means of the Gaussian distributions of said mixture 15'GM mixture having a height less than a first predefined threshold value.Figure 5 illustrates an advantageous embodiment of a process 200 according to the invention, particularly when the maxima(s) of a color profile are not extremely pronounced or significant with respect to the residual color intensity, i.e., a non-zero RCI intensity described on average by the detection zone 13 in the absence of analyte or between two color profiles. Thus, in Figure 5, we find a first color profile 15', referred to as the "control" profile, whose maxima 15'Ma, 15'Mb are intense (on the order of eight to twelve times the average residual RCI intensity), unlike the test or measurement color profile 15" whose maximum 15"M is weaker, i.e., on the order of one to two times the residual RCI color intensity. Such residual intensity results from suboptimal capture conditions.Calculating the 115' color intensity as the area under the curve 100 for the abscissa range 15'ra is acceptable because it introduces a small approximation error, as the distribution of color intensities that characterizes it is much more significant than the residual color intensity RCI. However, performing such a calculation of the 115" color intensity for the second 15" color profile would lead to a significant approximation. To prevent this problem, substep 233 of the calculation of the 115', 115" color intensity of an identified color profile includes a step of subtracting from the area under the intensity curve 100 over the determined abscissa range 15'ar, 15"ar characterizing a 15', 15" color profile, an area 15'ria, 15"ria of residual color intensities.Such an area 15'ria, 15”ria of residual color intensities of a color profile 15', 15” can consist of an area under a curve 101, 102 defined by an affine function linking the beginning 15'b, 15”b and the end 15'e, 15”e of the color profile 15', 15”, on said range of abscissas 15'ar, 15”ar characterizing said color profile 15', 15”.

[0094] In this way, a process 200 according to the invention estimates more accurately the color intensities 115i, in this case in figure 5, the intensities 115' and 115”, of each color profile 15i identified in substep 232, independently of the presence or absence of a residual RCI color intensity.

[0095] In the presence of a first 15' and a second 15" color profile identified in the working image 20s' which digitally transcribes the development zone 13 of a reactive strip 1, a process 200 according to the invention can implement a step 240 of producing a measurement M) of an analyte of interest from the respective color intensities 115', 115" of two color profiles 15', 15" expressed in a digital representation 20s' of the development zone 13 of said reactive strip 1.

[0096] According to a preferred embodiment, such a step 140 of producing a measurement M of an analyte of interest can be derived from calculating a ratio between the respective color intensities 115' and 115" of two color profiles 15' and 15". Thus, the presence of such an analyte is determined and measured as a percentage with respect to the reactivity potential of said test strip 1, expressed by the first color profile, referred to as the "control color profile". The higher this ratio between the color intensities 115' and 115", respectively of the control profile and the measurement profile 15", the more the analyte concentration is confirmed.

[0097] Quantification or measurement of the analyte of interest in the liquid biological sample received by the test strip 1 can be obtained in step 240 by comparing the ratio of the respective color intensities of the two color profiles to predefined reference data. This comparison thus translates the ratio of intensities to the concentration of the analyte in the liquid biological sample received by the test strip 1. The reference data can be pre-established from the measurement of the analyte of interest at several concentrations in the liquid biological sample. Such reference data can be in the form of a calibration curve (concentration - ratio of intensities) associated with a test or a specific type T1 of test strip, for example, determined in an optional substep 223 of a process 200 according to the invention.A mobile terminal 2 can then deliver said measurement M via its output human-machine interface 25 to the user U. This requires that said mobile terminal 2 has been previously adapted, by prior installation in program memory 23, of a computer program product whose execution of program instructions by the processing unit 21 triggers the implementation of a measurement method 200 according to the invention. For this purpose, step 240 of a method 200 according to the invention can cause the output of said measurement M of an analyte of interest via the output human-machine interface 25 of the mobile terminal 2 as soon as it is produced.The user U of such a mobile terminal 2 thus has on the fly, i.e. in near real time, after a digital representation 20s of the reactive strip 1 has been captured in a step 210 of said process 200 by the image sensor 22 of his mobile terminal 2, then selected and a measurement M of the concentration of an analyte of interest has been produced in step 240 of such a process from respective colour intensities of at least two colour profiles identified in a step 230 by automatically exploiting a colour distribution expressed in a working image 20s' produced in a step 220.As an alternative or in addition to such a display, or more generally such direct feedback to the user U, said step 240 of a method 200 according to the invention may cause the measurement M and / or the ratio of color intensities, or even the color intensities of the identified profiles, to be recorded in the data memory 24 of the mobile terminal 2. Such a step 240 may also, or alternatively, cause the measurement to be transmitted to a remote application server 3 via communication means 26 of said mobile terminal 2.

Claims

DEMANDS 1. A method for the quantitative measurement by chromatography of an analyte of interest, said method being designed to be implemented by a processing unit (21) of a terminal (2) comprising an image sensor (22) and an output human-machine interface (25), said method (200) comprising: - a step (230) of producing a colour intensity (115i, 115', 115”) for each colour profile (15', 15”) expressed in a digital representation (20s') of a development zone (13) of a reactive strip (1); - a step (240) of producing a measurement (M) of an analyte of interest from the respective color intensities (115', 115") of two color profiles (15', 15") expressed in said digital representation (20s') of the development zone (13) of said test strip (1); said method (200) being characterized in that it is arranged to be fully implemented by the processing unit of a "consumer" terminal (2) in the absence of specific devices for illuminating the test strip (1) and / or positioning the latter (1) with respect to the image sensor (22) of said terminal (2) during the acquisition of the digital representation (20s') of the development zone (13) of the test strip (1), said arrangement consisting in that: - it includes: o a step (210) of acquiring a stream of digital representations (20i) of the reactive strip (1) delivered by the image sensor (22) and of selecting on the fly one of said digital representations (20i) as soon as it satisfies one or more relevance criteria; o a step (220) of developing a working image (20s') from said selected digital representation (20s), encoding a light intensity per pixel, said working image describing a distribution of color intensities (100) of said development area (13) of said reactive strip (1) in the form of a distribution of light intensities of the pixels forming said working image (20s'), said step (220) of developing a working image (20s') comprising: ■ a substep (221) of extracting a region of interest within said selected digital representation (20s) including said revealing area (13) of the reactive strip (1); ■ a substep (222) of correcting geometric distortions in the selected digital representation (20s) with regard to an expected shape of a typical development zone (13) of a reactive strip (1), said geometric distortions arising from capture incidents during the step (210) of acquiring a stream of digital representations (20i); - the step (230) of producing a color intensity (115i, 115', 115”) for each color profile (15', 15”) expressed in a digital representation (20s') of a development area (13) comprises: o a substep (231) of modeling the distribution of color intensities expressed in the working image (20s') in the form of a color intensity curve (100) describing the cumulative light intensity by each set of pixels aligned perpendicularly to a reading axis (AL) of the revelation area (13); o a substep (232) of identifying a color profile (15', 15”) from said color intensity curve and of determining a range of abscissas (15'ar, 15”ar) characterizing the beginning (15'b, 15”b) and the end (15'e, 15”e) of said color profile (15', 15”); o a substep (233) of calculating the color intensity of said identified color profile (15', 15”) as being the area under the intensity curve (100) on said determined range of abscissas (15'ar, 15”ar); - the step (140) of producing a measurement (M) of an analyte of interest: o consists of calculating a ratio between the respective colour intensities (115', 115”) of two colour profiles (15', 15”); o causes an output of said measurement (M) of an analyte of interest produced by the output human-machine interface (25) of the terminal (2).

2. A method (200) according to claim 1, wherein an instance of the substep (232) of identifying a color profile (15', 15”) from said color intensity curve and determining a range of abscissas (15'ar, 15”ar) characterizing the beginning (15'b, 15”b) and the end (15'e, 15”e) of said color profile (15', 15”), as well as an instance of the substep (233) of calculating the color intensity of said identified color profile, are iterated (234) from the lowest abscissa towards the highest abscissa of said color intensity curve (100).

3. Method (200) according to claim 1 or 2, wherein the substep (233) of calculating the intensity (115', 115”) of colouring of said identified colouring profile comprises a subtraction step from the area under the intensity curve (100) on said determined abscissa range (15'ar, 15”ar) characterizing a colouring profile, of an area (15'ria, 15”ria) of residual colouring intensities.

4. Method (200) according to claim 3, wherein the area (15'ria, 15”ria) of residual colour intensities of a colour profile (15', 15”) consists of the area under a curve (101, 102) defined by an affine function linking the beginning (15'b, 15”b) and the end (15'e, 15”e) of the colour profile (15', 15”) on said range of abscissas (15'ar, 15”ar) characterizing said colour profile (15', 15”).

5. A method (200) according to any one of the preceding claims, wherein the step (210) of acquiring a stream of digital representations (20i) of the reactive strip (1) delivered by the image sensor (22) and of selecting, on the fly, one of said digital representations (20i) comprises: - a substep of capturing (201) a digital representation (20i); - a substep (203) of searching within said digital representation (20i) for a contour of an internal surface (12) of a reactive strip (1) comprising a development zone (13), and of selecting said digital representation (20i): o if a surface contour (12) describing a contour of a predefined geometric shape of reactive strip surface (1) has been identified, and o if a ratio between the area captured by such a surface contour (12) and the total area of ​​said digital representation (20i) is greater than and at a predetermined threshold.

6. A method according to the preceding claim, wherein the step (210) of acquiring a stream of digital representations (20i) of the reactive strip (1) delivered by the image sensor (22) and of selecting, on the fly, one of said digital representations (20i) comprises a substep of displaying it by the output human-machine interface (25) of the terminal (2) such that a suitable geometric shape describing a typical template of a "control" reactive strip is superimposed transparently on said digital representation (20i) in order to induce the user (U) of the terminal (2) to position the image sensor (22) of the terminal (2) relative to the reactive strip (1) so that said suitable geometric shape covers the visible surface of the reactive strip (1) on said digital representation (20i) displayed by said output human-machine interface.

7. Method (200) according to any one of the preceding claims, wherein the relevance criterion or criteria that a digital representation (20i) must satisfy in order to be selected on the fly (203) is or are chosen from a set of criteria including ambient brightness during capture, sharpness, contrast, color temperature of the captured digital representation (20i) and the content that the latter expresses.

8. A method (200) according to any one of the preceding claims, wherein the step (220) of developing a working image (20s') from said selected digital representation (20s) comprises a substep (223) of determining the type (Tl) of chromatographic test associated with the reagent strip (1) by implementing a machine learning algorithm, and of configuring: - the step (230) of producing a colour intensity (115i, 115', 115”) for each colour profile (15', 15”) expressed in the digital representation (20s') of the development zone (13); - the step (140) of producing a measure (M) of an analyte of interest.

9. Product computer program (P) comprising program instructions which, when written into a program memory (23) of a terminal (2) further comprising a processing unit (21), an output human-machine interface (25) and a data memory (24), and interpreted or executed by said processing unit (21) of the latter (2), cause the implementation of a method (200) according to any one of claims 1 to 8.

10. Mobile terminal (2) comprising a program memory (23), a processing unit (21), an output human-machine interface (25) and a data memory (24), characterized in that the program memory (23) contains the program instructions of a computer program product according to claim 9.

Images