Method for determining the susceptibility of a microorganism to an antimicrobial agent on the basis of a spectral image

Abstract

The invention relates to a computer-implemented method for characterising a microbial strain, which method comprises the following steps: - obtaining a spectral image of a colony, representing a colony of a microbial strain, the spectral image of the colony comprising a plurality of pixels, each pixel comprising a plurality of spectral channels; - filtering the spectral image of the colony in order to obtain a "test image", each pixel in the test image being dependent on an average of a packet of neighbouring pixels in the image of the colony, each packet comprising M<N pixels, where N is the number of pixels in the image of the colony; - classifying each pixel in the test image predicting whether the spectrum of the pixel belongs to a class of a prediction model having at least two microbial strain characterisation classes; - determining the class characterising the microbial strain by a majority vote on the set of predicted classes for each pixel; - determining a susceptibility of the microbial strain to a microbial agent as being that associated with the class characterising the microbial strain.

Description

[0001] DESCRIPTION

[0002] TITLE OF THE INVENTION: Method for determining the susceptibility of a microorganism to an antimicrobial agent from a spectral image

[0003] TECHNICAL FIELD

[0004] The invention relates to the field of microbiological analysis, and in particular to the characterization of microorganisms, notably the prediction of the sensitivity or resistance of yeasts, molds, and bacteria to an antimicrobial agent. More specifically, the invention applies to the analysis of a spectral image of one or more colonies of bacteria, molds, or yeasts grown in or on an observable culture medium.

[0005] STATE OF THE ART

[0006] In the field of in vitro diagnostics of microorganisms, particularly pathogens, characterizing a microorganism primarily involves identifying its species and its susceptibility to an antimicrobial agent (or "antibiogram") in order to determine a treatment for a patient infected by that microorganism. To achieve this, a complex microbiological process is typically implemented in the laboratory, a process that most often requires prior knowledge of other properties of the microorganism, notably its kingdom (e.g., yeast or bacteria), and, in the case of bacteria, its Gram stain type or whether it is fermentable or not. Such a process includes a significant number of manual steps (for example, for Gram stain determination: fixation, staining, mordanting, washing, overstaining, etc.), and is therefore time-consuming.

[0007] To reduce human intervention, spectral imaging in the visible and near-infrared has been used, decreasing the number of samples and mass spectrometry or biomolecular analyses downstream of the culture stage by classifying colonies into distinct groups, particularly at the species level, or at lower levels of classification in the phylogenetic tree, notably for typing or detecting these clonal groups. Documents WO2019 / 122732 and WO2023 / 094775 describe such processes. Hyperspectral and multispectral imaging in the visible and near-infrared is known to enable microbial characterization on Petri dishes with correct classification rates exceeding 90%. Given that microbial characterization is a highly sensitive application, particularly in terms of human, animal, and environmental health, it is essential to achieve ever-improving performance.

[0008] DESCRIPTION OF THE INVENTION

[0009] The invention makes it possible to predict the susceptibility of a microorganism to an antimicrobial agent using hyperspectral or multispectral imaging of a microbial colony grown on a culture medium without the presence of said antimicrobial.

[0010] To this end, the invention proposes a computer-implemented method for characterizing a microbial strain, comprising the following steps:

[0011] - obtaining a spectral image of a colony, representing a colony of a microbial strain, the spectral image of the colony comprising several pixels, each pixel comprising several spectral channels;

[0012] - filtering of the spectral image of the colony to obtain a "test image", each pixel of the test image being a function of an average of a packet of neighboring pixels from the colony image, each packet comprising M <N pixels avec N le nombre de pixels de l’image de la colonie ;

[0013] - classification of each pixel of the test image predicting whether the pixel spectrum belongs to a class of a prediction model with at least two classes of microbial strain characterization;

[0014] - determination of the class characterizing the microbial strain by a majority vote on all the classes predicted for each pixel;

[0015] - determination of a susceptibility of the microbial strain to a microbial agent as being that associated with the class characterizing the microbial strain.

[0016] According to one embodiment, the filtering consists of grouping the N pixels of the colony image into packets of M <N pixels voisins, un pixel de l’image de test étant la moyenne des M pixels voisins de l’image de la colonie, l’image de test présentant un nombre de pixels inférieurs à celui de l’image de la colonie.

[0017] In one embodiment, the filtering process consists of applying a convolutional filter to the colony image, where a pixel in the test image is obtained by processing several neighboring pixels in the colony image using a convolutional kernel. The test image comprises N pixels, each pixel corresponding to an average of M <N pixels voisins de l’image de la colonie, l’image de test présentant un nombre de pixels identique à celui de l’image de la colonie.

[0018] According to one embodiment, M is between 2 and 10, preferably equal to 5.

[0019] According to one embodiment, the threshold for the majority vote is between 60% and 80%, in particular 60% or 80%.

[0020] According to one embodiment, a prediction model is a model with at least three classes, each class corresponding to a microbial strain of a type, each strain being previously characterized as susceptible or not to a microbial agent, the susceptibility or resistance of the microbial strain to a microbial agent being that associated with the microbial class.

[0021] According to one embodiment, the prediction model is a 2-class model, each class indicating susceptibility or not to a microbial agent, the susceptibility of the microbial strain to a microbial agent being the result of the step of determining the class characterizing the microbial strain.

[0022] The method of the invention may include training the prediction model, the training comprising the following steps:

[0023] - obtaining spectral images of different colonies for several microbial strains;

[0024] - training in classification on spectral images of different colonies.

[0025] According to one embodiment, the spectral image comprises for each pixel 240 channels spaced 2.1nm apart over a band between 390 nm and 900 nm, the spectral image being hyperspectral.

[0026] According to one embodiment, the spectral image comprises for each pixel 23 channels spaced 10 nm apart over a spectral band between 390 nm and 900 nm, the image being multispectral.

[0027] According to one embodiment, the microbial strain is a strain of Staphylococcus aureus and the antimicrobial agent is methicillin.

[0028] According to one embodiment, the microbial strain is obtained from a sample taken from a patient, and the process comprises a step of selecting an antimicrobial therapy based on the determination of the microbial strain's susceptibility and administering said therapy to the patient. The invention also relates to a computer program product comprising code instructions for executing a process according to the invention.

[0029] The invention also relates to a method for learning a prediction model implemented in the characterization method according to the invention, comprising the following steps:

[0030] - creation of a learning database including, for each strain of a species, its susceptibility to the microbial agent as well as corresponding colony spectra;

[0031] - supervised training of the prediction model in order to obtain prediction model parameters on spectra of different colonies.

[0032] Filtering the spectral image improves classification performance, making it more robust.

[0033] Furthermore, spectral imaging, preferably between 390 nm and 900 nm, contains sufficient information to predict whether two microbial strains are clonal or derived from the same lineage and thus share the same susceptibility to the antimicrobial agent. By knowing the susceptibility of a class, and by predicting that a strain belongs to that class, the strain to be characterized can be predicted to have the same susceptibility to that class.

[0034] PRESENTATION OF THE FIGURES

[0035] Other features, purposes and advantages of the invention will become apparent from the following description, which is purely illustrative and not limiting, and which should be read in conjunction with the accompanying drawings on which:

[0036] - Figure 1 illustrates a system for characterizing a microbial strain according to the invention;

[0037] - Figure 2 illustrates a spectral imaging system according to a first embodiment;

[0038] - Figure 3 illustrates a spectral imaging system according to a second embodiment;

[0039] - Figure 4 illustrates steps in a process for characterizing a strain according to an embodiment of the invention; - Figure 5 illustrates obtaining a test image from an image of a colony according to a first variant of a filtering implemented in the characterization process according to the invention;

[0040] - Figure 6 illustrates obtaining a test image from an image of a colony according to a second variant of a filtering implemented in the characterization process according to the invention;

[0041] - Figure 7 shows strains used by prediction models.

[0042] Across all figures, similar elements bear identical references.

[0043] DETAILED DESCRIPTION

[0044] System

[0045] Figure 1 illustrates a system S for characterizing a microbial strain comprising one or more colony(ies) of yeast or bacteria grown on an agar poured into a Petri dish.

[0046] Such a system S comprises a device 1, T for acquiring spectral images of a microbial strain 2 preferably contained in a Petri dish 21 and a processing unit 3.

[0047] The processing unit 3 is connected to the spectral image acquisition device 1, T, either wired or wirelessly. It comprises one or more processors configured to optionally control the acquisition device, and one or more data storage units storing instructions to implement steps in a process for characterizing a microbial strain based on acquired spectral image(s) described below.

[0048] The processing unit 3 is also connected to one or more servers 4 which store one or more databases. A database can store spectral images acquired for processing, training data, classification data, etc.

[0049] The processing unit 3 can be associated with a user interface and a screen. Imaging device

[0050] In relation to Figure 2, the spectral imaging device 1 is, in a first embodiment, of the hyperspectral type. It advantageously comprises: a so-called hyperspectral camera 18, consisting of a digital sensor comprising an array of elementary sensors, for example a CCD or CMOS type digital sensor, sensitive in a range of wavelengths, for example [Â min ; HAS max ] = [400 nm; 900 nm]; and a light-dispersive element or a spectrograph to select a wavelength to be acquired by the sensor; a lens 20 to focus onto the digital sensor of the camera 18, the optical image of the strain 2 from which a hyperspectral image is to be acquired; a front illumination 24, for example consisting of one or more halogen lamps, for example two or four lamps, capable of emitting light in the range [2 min ; HAS max] and to achieve uniform front illumination of strain 2. For example, the lights are white light lamps; a rear light 26, for example consisting of a matrix of white light LEDs, to achieve, depending on the case, uniform rear illumination of strain 2 in the range; a carriage 28 on which rests the Petri dish 21 containing strain 2 and allowing the latter to pass in front of the lens 20 in order to obtain a whole image of it by scanning.

[0051] It is noted that the rear lighting is used with or instead of the front lighting to enable transmission acquisition from camera 18.

[0052] Device 1, for example, is configured to acquire the image of a 90 mm by 90 mm region with a sampling step of 160 micrometers (estimated spatial resolution of 300 micrometers) and with a spectral resolution of a few nanometers over the range [A min ;2 max], It is possible to exceed 200 channels over a range of approximately 500 nm. In particular, the field of view and depth of field of objective 20 are chosen for obtaining images that can include complete colonies with a radius of up to 1 cm, preferably up to 0.9 cm, and even more preferably 0.5 cm.

[0053] Device 1 thus produces a digital HSI (Hyper Spectral Imaging) image of the light reflected by sample 2, improperly called a "hypercube" because it is in fact three-dimensional: two spatial dimensions and one spectral dimension, each pixel (or rather voxel due to the three-dimensional nature of the HSI image) representing the radiance measured at a point of sample 22 for a spectral channel. A pixel thus contains a spectrum.

[0054] The radiance of a pixel, commonly called "light intensity", corresponds here to the amount of light incident on the surface of the corresponding elementary sensitive site of the camera sensor 18 during the exposure time, as is known in itself from the field of digital photography for example.

[0055] In relation to Figure 3, the spectral imaging device T is in a second embodiment of the multispectral type and comprises a camera 34, advantageously a high spatial resolution CMOS or CCD camera, coupled to an array of spectral filters 36, for example arranged in front of the lens 20 between the lens 20 and the sensor of the camera 32. The array of filters 36 consists of a number NF of distinct bandpass filters, each configured to transmit only light in a part of the range [A min ;2 max], with a spectral width at half maximum (FWHM) less than or equal to 10 nm, 25 nm, 40 nm or 50 nm, and preferably less than or equal to 20 nm. Assembly 36 is, for example, a filter wheel that can accommodate several different filters, a wheel driven by the data processing unit which actuates it to scroll in front of the camera and trigger an image capture for each of them.

[0056] Device T thus produces a digital MSI (Multi Spectral Imaging) image of the light reflected by sample 2.

[0057] Characterization process

[0058] A method for characterizing a microbial strain is described in relation to Figure 4.

[0059] In a step of strain preparation (step E1), a Petri dish 21 is inoculated with a biological sample, for example, taken from a patient (urine, blood, or more generally any biological fluid), in order to grow bacterial colonies on the surface of a nutrient medium, or "culture medium," placed in the Petri dish. The nutrient medium's primary purpose is to promote colony growth and, optionally, to enhance the accuracy of the characterization by limiting light interference. This type of growth is well known to those skilled in the art and will not be described in further detail.

[0060] Advantageously, the preparation is carried out so that there is at least one area of ​​the Petri dish on which the colonies are spaced apart from each other to be easily separable on the image that will be acquired and therefore their segmentation.

[0061] Once colony growth is complete, for example after 16, 24, 36, or 48 hours, the Petri dish is opened for analysis. This is referred to as the microbial strain to be characterized.

[0062] Box 2 is then positioned on the carriage 28 of the imaging device (step E2) and at least one spectral image of the box is acquired 2 and stored in processing unit 3 (step E3).

[0063] Depending on the case, we have a hyperspectral image l_HSI acquired with the imaging device 1 of figure 2 or a multispectral image l_MSI acquired with the imaging device 1' of figure 3.

[0064] The acquired image is advantageously pre-processed to filter fluctuations in the emission spectrum of the lighting sources 24, 26. For example, a "flat field correction" (FFC) type correction is implemented to obtain the reflectance, which also has the advantage of correcting pixel-to-pixel sensor response dispersions (step E4).

[0065] In the case of hyperspectral imaging, each pixel of the image, and therefore each spectrum, comprises 240 channels across the wavelength range between 400 nm and 900 nm, with a step size of 2.1 nm. This resolution is considered high, which then allows, through grouping, the simulation of the spectral width of commercially available color filters. For example, the 10 nm width of a multispectral image is simulated by averaging three contiguous channels of a hyperspectral image.

[0066] In the case of multispectral imaging, each pixel comprises several channels with widths of 10, 25, 40, or 50 nm in 10 nm increments. Preferably, the width is 10 nm. For a full width at half maximum (FWHM) of 10 nm, each pixel comprises 48 channels. For FWHMs of 25 nm and 50 nm, each pixel comprises 18 channels. Bacterial colony identification and segmentation are then performed (step E5). For example, a bacterial colony identification algorithm is implemented on the l_HSI or l_MSI images. Any conventional pattern and object recognition algorithm can be used to extract an area from the l_HSI or l_MSI image, labeled Col_HSI or l_MSI, corresponding to a colony. Alternatively, this selection can be made manually by an operator who selects this area in the image using the display screen and a user interface.In general, segmentation allows the detection of all colonies of interest, removing artifacts such as filaments or dust. Segmentation can be implemented in any known way. Alternatively, an operator selects, for example using a graphical interface, one or more colonies to be characterized.

[0067] It should be noted that each pixel of the spectral image comprises a spectrum, which is a curve representing the light intensity measured at the colony scale as a function of frequency. This is a vector whose size is the number of channels in the spectral image. As an example, with the acquisition device described above, a colony that can be circular in shape, with a diameter ranging from a few tenths of a millimeter to a few millimeters, can extend over an area of ​​the spectral image from ten to several hundred pixels. In cases where the spectral image represents several colonies of the same strain, the procedure described here applies to each colony or to a group of colonies chosen according to criteria such as size or position in the culture medium.

[0068] Advantageously, segmented image smoothing is implemented using conventional techniques (step E6). Techniques known by their English names, such as Raw, Smoothing, SNV, First derivative, or Second derivative, can be applied.

[0069] Once a colony is identified, a processing (step E7) of the spectral image of the colony is implemented.

[0070] According to a first implementation, such processing consists of applying convolutional filtering to the spectral image of the colony. The filtered image is the same size as the colony image. This filtering reduces noise in the spectral image. For example, each pixel in the filtered image is the average of a group of neighboring pixels in the colony's spectral image. Thus, this filtering groups the pixels of the colony image in a sliding fashion. In this first implementation, between 1 and 9 neighboring pixels are considered, preferably 4 neighboring pixels. Figure 5 illustrates such a first implementation. On the left is the original image of the colony, and on the right is the filtered image (test image). A pixel in the test image corresponds to the average of 4 neighboring pixels in the colony image. The 4 neighboring pixels are grouped in a sliding fashion.

[0071] According to a second implementation, such processing also involves grouping neighboring pixels of the colony image into pixel packets. Each pixel of the filtered image then corresponds to the average of the pixels in each packet. In this case, the filtered image is smaller than the colony image. This second implementation involves grouping neighboring pixels into packets of 2, 3, 4, or 5, preferably 5, and then averaging the spectra of the selected pixels. The filtered image then comprises pixels that are the average of several pixels from the spectral or hyperspectral image. By grouping the pixels, noise is also reduced. Figure 6 illustrates such a second implementation. On the left is the colony image (i), and on the right is the filtered image (tit, or test image). One pixel in the test image corresponds to the average of four neighboring pixels in the colony image.Here there is no sliding grouping, the pixel packets are neighbors and independent.

[0072] With the processing according to the first or second realization we obtain a "test image", each pixel being a function of the pixels of the colony image and preferably of the average of several neighboring pixels.

[0073] It should be noted that, according to the prior art of microbial characterization by hyperspectral or multispectral imaging, for example described in applications WO2019 / 122732 and WO2023 / 094775, the spectra of the pixels of a colony are averaged, thus producing a single spectrum per colony, which is then processed to obtain the desired information. All pixels are therefore averaged. In another variant, each pixel of the colony, and therefore each spectrum associated with a colony, is processed to obtain this information, and a final consensus is reached by majority vote. To improve the performance of these prior art processes based on colony averaging or pixel-by-pixel processing followed by majority vote, the aim was to improve training sets (particularly in terms of diversity, data quantity, cross-validation strategy, etc.).) and on the application of other algorithms, in particular those based on deep learning (notably convolutional neural networks).

[0074] In the invention described herein, the inventors do not seek to improve training databases or predictive models, but rather to characterize the signal-to-noise ratio of spectra. Indeed, the inventors observed that averaging spectra across a colony, and therefore averaging several dozen, or even a hundred, spectra, results in the loss of important information for characterizing microorganisms. Furthermore, the spectrum of a single pixel is so noisy that this "fine" information is difficult to exploit. The inventors thus found that averaging 2 to 10 spectra reduces noise while preserving this information, with the best performance observed being associated with averaging 3 to 7 spectra, and the peak performance occurring at 5 spectra.

[0075] This shows the advantage of grouping the pixels of the colony image into several packets so that the pixels of the test image are a function of an average of these packets.

[0076] The process then continues with the classification of each pixel in the test image, predicting whether the pixel's spectrum belongs to a class within a prediction model with at least two classes characterizing the microbial strain (step E8). Following classification, each pixel is either assigned to a class or remains unassigned.

[0077] Next, the class characterizing the microbial strain is determined by a majority vote across all predicted classes for each pixel (step E9). This involves considering all classes and calculating which class stands out. The strain's class is the one with the most votes. Class determination is effective for a vote percentage between 60% and 80%, particularly for a 60% or 80% vote percentage.

[0078] Based on the strain class, the susceptibility of the microbial strain to a microbial agent is determined as that associated with the class characterizing the microbial strain (step E10). Indeed, each class is associated with susceptibility or non-susceptibility to a microbial agent.

[0079] Multi-class model

[0080] In one embodiment, the prediction model allows the spectrum of a pixel to be predicted to which strain, among several strains, each strain being either susceptible or not to a microbial agent. This is referred to as a multi-class model (more than two classes). In the case of the Staphylococcus aureus strain, the prediction model includes 73 strains or classes that are either susceptible or not to methicillin. After classification, each pixel is associated with a strain or is left unassigned. Following the vote, a Staphylococcus aureus strain is obtained, and its susceptibility to methicillin is associated with this strain in the database. For example, MRSA stands for "Methicillin-resistant Staphylococcus aureus" and MSSA for "Methicillin-sensitive Staphylococcus aureus," that is, Staphylococcus aureus strains that are respectively resistant and not resistant to the methicillin antibiotic.

[0081] As such, the model uses a database comprising 73 strains of Staphylococcus aureus, each associated with spectral images of colonies obtained under the same conditions as the characterization procedure. Each strain is associated with an MSA or MRSA susceptibility. The model provides the strain here.

[0082] According to one embodiment, the prediction model makes it possible to predict the microbial species to which the colony belongs.

[0083] 2-class model

[0084] In one embodiment, the prediction model allows us to predict whether a pixel belongs to a strain susceptible or not to a microbial agent. This is referred to as a two-class model. In the case of the Staphylococcus aureus strain, the prediction model comprises two classes: MRSA or MSA. Following classification, each pixel is either associated with a susceptibility or is unassigned. Majority voting determines whether the strain is MRSA or MSA.

[0085] As such, the model uses a database comprising 73 strains of Staphylococcus aureus, each associated with spectral images of colonies obtained under the same conditions as the characterization procedure. Each strain is associated with an MSA or MRSA susceptibility. The model directly provides the methicillin susceptibility.

[0086] The model, whether with two or more classes, is a support vector classification (SVM) model or a convolutional neural network (CNN). In the case of an SVM, one might choose, for example, an RBF kernel SVM (Radial Basis Function).

[0087] Learning

[0088] The process advantageously includes a learning step (step E0) using a training database stored in the processing unit 4.

[0089] The training database for the microbial species under consideration is constituted as follows (step E01).

[0090] For each strain of a species, the strain is prepared in a Petri dish as in step E1 described previously.

[0091] Specifically, for each strain, several samples (e.g., three, preferably four) are prepared, and spectral images are acquired for each colony in each sample using device 1, T as described in Figures 2 and 3, preferably from several colonies per Petri dish. The spectrum is processed as in step E4. The spectra of the first two samples are used for training microbial strain classes, while the spectra of the remaining samples (e.g., the third sample in the case of triplet production) are used to test the performance of the training. Optionally, acquisition is also performed on several samples using different capture devices to capture spectral variability caused by differences in device characteristics (e.g., variability in light sources between devices, etc.).

[0092] The phenotypic measurement of the strain's susceptibility to the antimicrobial agent, and its memorization, is carried out.

[0093] Finally, a genomic characterization of the strain is carried out using a complete sequencing of its genome and the establishment of a wgMLST profile as described in the document "MLST revisited: the gene-by-gene approach to bacterial genomics" by Martin CJ Maiden, Nature Reviews Microbiology, 2013, and the recording of this characterization.

[0094] The learning database then includes, for each strain of a species, its susceptibility to the microbial agent as well as corresponding colony spectra.

[0095] In the case of Staphylococcus aureus, 73 different strains are used for training, each associated with methicillin susceptibility or non-susceptibility. This is shown in Figure 5.

[0096] Once the database is established, the training of the prediction model itself (step E02) is implemented in a conventional manner.

[0097] Example

[0098] A prediction model typically consists of two parts: the first part produces a score for each class (for example, between 0 and 1 for a multi-class prediction), possibly associated with a confidence level; and the second part applies a predetermined rule to these scores to produce a final classification result, such as selecting the class associated with the highest score or selecting the classes whose scores exceed a predetermined threshold. In this second case, for each pixel followed by a majority vote, a pixel is unassigned when none of the scores exceeds the single threshold, or, when a confidence level is generated, when the single score does not exceed a predetermined threshold. Similarly, when the majority of pixels do not exceed the voting threshold, the colony is assigned the unassigned status.

[0099] For treatments based on the average spectrum of a colony, the colony is assigned the status of unassigned when none of the scores exceeds a threshold or, if applicable, the confidence level is too low. Regardless of the algorithm used, how it predicts classes, and / or calculates a confidence level, it is important to note that for sensitive applications, such as the in vitro diagnosis of microorganisms infecting a patient, it is crucial to have the highest possible specificity for positive results, even at the expense of sensitivity (and therefore the "unassigned" prediction). Indeed, false positives or true negatives can have detrimental consequences for the patient's health, such as choosing an ineffective antibiotic based on a prediction of a bacterium's susceptibility to that antibiotic.When a prediction returns an "unassigned" result, microbiologists can then launch additional tests to obtain the desired information.

[0100] Thus, improving a predictive model applied to in vitro diagnosis involves, in particular, being demanding about the certainty of the prediction while limiting as much as possible the non-assignment rate which requires carrying out other tests.

[0101] The invention has been applied to the characterization of methicillin-resistant Staphylococcus aureus microbial strains in order to identify MRSA and MSA strains. Figure 7 shows MRSA and MSA strains used.

[0102] Several levels were tested:

[0103] Colony: classification according to the whole colony by averaging the spectra of the pixels and direct classification by a classification model.

[0104] Pixel: classification of each pixel by the classification model.

[0105] Packets: grouping into packets of 5 pixels and using spectra for each packet and classification of each packet by the classification model.

[0106] In what follows, the classification rate corresponds to a correct classification, while not assigned means that the classification did not return a result, as explained above.

[0107] Scenario 1: 73-class model - Hyperspectral image

[0108] The best performance was achieved with the pixel level and a majority vote of 80%. However, 81.7% were unassigned.

[0109] The packet level made it possible to reduce the percentage of unassigned to 43.7% with a very satisfactory classification rate.

[0110] Scenario 2: MRSA / MSA 2-Class Model - Hyperspectral Image

[0111] The best performance is clearly achieved at the packet level with a low percentage of unassigned data. Scenario 3 - 73-class model - comparison of hyperspectral HSI / multispectral MSI image (23 channels, 10 nm width)

[0112] Scenario 4 - 2-class model - comparison of hyperspectral HSI / multispectral MSI image (23 channels, 10 nm wide) Scenarios 3 and 4 demonstrate good performance when using the packet level, particularly with the 2-class model, by reducing the percentage of unassigned values ​​compared to the 73-class model. Using an MSI image also achieves good performance while reducing image acquisition complexity.

[0113] In summary, the best performance for phenotype identification is obtained at the package level and 80% majority vote with a classification rate of 97.1% and 15.3% of colonies not assigned.

[0114] Compared to the entire spectrum at the same level of analysis, all performances are significantly lower before the majority vote. The vote helps to improve performance.

[0115] Using an MSI image with a FWHM width of 10 nm leads to good performance.

Claims

DEMANDS 1. A computer-implemented method for characterizing a microbial strain, comprising the following steps: - obtaining (E3) a spectral image of a colony, representing a colony of a microbial strain, the spectral image of the colony comprising several pixels, each pixel comprising several spectral channels; - processing (E7) of filtering the spectral image of the colony to obtain a "test image", each pixel of the test image being a function of an average of a packet of neighboring pixels of the colony image, each packet comprising M <N pixels avec N le nombre de pixels de l’image de la colonie ; - classification (E8) of each pixel of the test image predicting whether the pixel spectrum belongs to a class of a prediction model with less than two classes of microbial strain characterization; - determination (E9) of the class characterizing the microbial strain by a majority vote on all the classes predicted for each pixel; - determination (E10) of a susceptibility of the microbial strain to a microbial agent as being that associated with the class characterizing the microbial strain.

2. A method according to claim 1, wherein the filtering consists of grouping the N pixels of the colony image into packets of M <N pixels voisins, un pixel de l’image de test étant la moyenne des M pixels voisins de l’image de la colonie, l’image de test présentant un nombre de pixels inférieurs à celui de l’image de la colonie.

3. A method according to claim 1, wherein the filtering process consists of applying a convolutional filter to the colony image, a pixel of the test image being obtained by processing several neighboring pixels of the colony image using a convolutional kernel, the test image comprising N pixels, each pixel corresponding to an average of M <N pixels voisins de l’image de la colonie, l’image de test présentant un nombre de pixels identique à celui de l’image de la colonie.

4. A method according to any one of claims 1 to 3 wherein M is between 2 and 10, preferably equal to 5.

5. A method according to any one of the preceding claims, wherein the threshold for the majority vote is between 60% and 80%, in particular 60% or 80%.

6. A method according to any one of claims 1 to 5, wherein the prediction model is a model with at least three classes, each class corresponding to a microbial strain of a type, each strain being previously characterized as susceptible or not to a microbial agent, the susceptibility or resistance of the microbial strain to a microbial agent being that associated with the microbial class.

7. A method according to any one of claims 1 to 5, wherein the prediction model is a 2-class model, each class indicating susceptibility or non-susceptibility to a microbial agent, the susceptibility of the microbial strain to a microbial agent being the result of the step of determining the class characterizing the microbial strain.

8. A method according to any one of claims 6 to 7, comprising a training (E0) of the prediction model, the training comprising the following steps: - obtaining spectral images of different colonies for several microbial strains; - training in classification on spectral images of different colonies.

9. A method according to any one of the preceding claims, wherein the spectral image comprises for each pixel 240 channels spaced 2.1nm apart over a band between 390 nm and 900 nm, the spectral image being hyperspectral.

10. A method according to any one of claims 1 to 8, wherein the spectral image comprises for each pixel 23 channels spaced 10 nm apart over a spectral band between 390 nm and 900 nm, the image being multispectral.

11. A method according to any one of the preceding claims, wherein the microbial strain is a strain of Staphylococcus aureus and the antimicrobial agent is methicillin.

12. A product computer program comprising code instructions for carrying out the steps of the method according to any one of the preceding claims when said program is executed on a computer.

13. A computer-based method for learning a prediction model implemented in the characterization method according to any one of claims 1 to 11, comprising the following steps: - constitution (E01) of a training database including for each strain of a species, its susceptibility to the microbial agent as well as corresponding colony spectra; - supervised training (E02) of the prediction model in order to obtain prediction model parameters on different colony spectra.

Images