Device for detecting bacteria in a sample

Abstract

The invention relates to a device for detecting bacteria in a sample, comprising an optical unit (4) which is arranged to produce sets of multispectral images of portions of a sample, obtained by illumination eliciting an autofluorescence response and by illumination eliciting a reflectance response. These sets of multispectral images are supplied to an analyser (6) comprising a classifier which produces an image in which each pixel is drawn from an identical pixel of the images of the set of multispectral images and is associated with an element identifier indicating whether this pixel is associated with a bacterium or with another element. This image is then used by a computer to determine bacteria detection data.

Description

Description Title of the invention: Device for detecting bacteria in a sample

[0001] The invention relates to the field of microbiology and in particular the detection of bacteria in samples.

[0002] In the food industry, bacteria detection is a major challenge. All stakeholders are subject to extremely strict hygiene standards to prevent the sale of food containing bacteria, regardless of its degree of processing. Similar issues are found in the pharmaceutical and cosmetic industries.

[0003] Indeed, when the presence of bacteria is detected too late, it is common for many consumers to suffer the consequences, which can range from more or less serious digestive discomfort to poisoning and / or infections that can cause death. A. State of the art

[0004] The state of the art in detecting the presence of bacteria is based on taking samples at regular intervals on production lines and counting the bacteria present in the samples after cultivation.

[0005] This conventional approach involves culturing bacteria for several days on a nutrient medium, followed by manual counting of bacterial colonies to determine the number of bacteria, usually measured in colony-forming units per milliliter (CFU / mL). However, although considered the gold standard in microbiology, this method has several drawbacks, including the significant time required to obtain results, which can range from two to five days depending on the specific microorganisms to be quantified. This time is even longer in the case of sterility testing, which can take up to 15 days in the pharmaceutical and cosmetic fields.

[0006] In recent years, alternative methods have been explored to try to replace the Petri dish culture technique: polymerase chain reaction (PCR), the most probable number approach (MPN), ATP measurement (ATP-metry), flow or solid-phase cytometry, methods based on microscopy, methods based on the analysis of hyperspectral measurements, or the use of excitation and emission matrices (Excitation and Emission Matrices), etc. Al PCR Methods

[0007] Among these exploratory methods, PCR offers the fastest results, usually in less than an hour. However, PCR has limitations because it can only detect specific genetic sequences (such as the 16S rRNA gene) like the QuickScan quantitative test (registered trademark). However, this gene is not present in all bacterial species, which can introduce bias. Generally speaking, each PCR test focuses on a specific genetic sequence, which makes these tests insufficiently "generalist". In addition, the use of reagents in these methods makes them expensive and less accessible for routine use compared to other methods. A.2 "Most probable number" method

[0008] The most likely number-based approach is a culture-based approach. Biomérieux has optimized this approach using microfluidics and coupling it to a fluorescent dye to develop a solution called Tempo (registered trademark). To test the suspension, it is loaded and incubated in a micro-chamber with a culture medium coupled with a fluorescent dye. The limit of detection (LOD) of this method is 1 UEC / mL (see the article by Cayer et al. "Evaluation of the Tempo® System: Improving the Microbiological Quality Monitoring of Human Milk", Erontiers in Pediatrics, Volume 8, 2020, https: / / doi.org / 10.3389 / fped.2020.00494) and a dynamic range of 3.7 log, allowing the detection of the total number of viable bacteria ("Total Viable Count" or "TVC") between 24 and 48 hours.

[0009] This detection time is too long and does not constitute sufficient progress compared to Petri dish culture. A.3 Methods based on ATP-metry

[0010] ATP-metry involves amplifying the signal generated by the presence of adenosine triphosphate (ATP). This allows for rapid results in seconds, but the method is highly susceptible to organic contamination from chemicals or detergents, which can lead to many false positives. Because ATP-metry is easy to deploy in the field and relatively inexpensive, it is frequently used, but it remains unsuitable as a single, reliable measure of bacterial presence. A.4 Flow cytometry methods

[0011] This method relies on illuminating a capillary containing the sample with a laser. It can use dyes to label the bacteria, and an optical sensor is used to collect the photons emitted by the labeled bacteria.

[0012] Flow cytometry still has significant limitations. First, it counts both live and dead cells and requires the analysis of large sample volumes, which can take several hours and is too time-consuming. Furthermore, it is only compatible with liquid matrices and the equipment is expensive (several hundred thousand euros per machine). A.5 Solid-phase cytometry methods

[0013] The solution to be analyzed is filtered through a membrane selected to retain elements at least the size of a bacterium (e.g. 0.4 μm). This is combined with labeling the bacteria with a dye. Then, the membrane is scanned by a laser to excite the fluorochrome, and the number of bacteria is counted. As an alternative to laser detection, epifluorescence imaging can be performed.

[0014] Various implementations based on this principle have been developed by companies such as Red Berries (see for example patent application EP3861125A1), Microbs, Innosieve Diagnostics (see for example patent application EP2440331A1) or Mibic.

[0015] These methods require filterable solutions and expensive reagents. Furthermore, after detecting and counting microorganisms, the reagents used can have toxic effects on cell metabolism, which can hamper subsequent cultivation and identification of these microorganisms.

[0016] Finally, these methods require significant pretreatment steps, and incubation periods of at least 20 minutes, and they have a high level of false negatives due to dust or impurities affecting fluorescence, or dye absorption. A.6 Optical microscopy methods

[0017] Optical microscopy methods have attracted interest for the identification and enumeration of bacteria due to their potential for rapid, non-destructive detection that requires minimal sample preparation.

[0018] Characterization and enumeration of bacteria are required for a wide range of concentrations and different sample types, which significantly complicates these methods.

[0019] Current optical methods applied to the detection of bacteria only work with a very concentrated suspension, which makes them useless in practical applications. A.7 Epifluorescence microscopy methods

[0020] Fluorescence microscopy is an imaging technique that occurs when a fluorophore interacts with light of chosen energy, which leads to the excitation of a fluorophore from the ground state to a higher energy state, before returning to a lower energy level by emitting a photon.

[0021] Epifluorescence microscopy typically uses a single wavelength band and does not offer the ability to differentiate bacteria from other objects without the use of optimized bandpass filters. Indeed, the multiband approach requires switching between bands with a dichroic wheel, which is slow.

[0022] These methods additionally require sample preparation with optical dyes, such as DAPI, for bacterial counting (see the article by T. Muthukrishnan et al. "Evaluating the Reliability of Counting Bacteria Using Epifluorescence Microscopy", Journal of Marine Science and Engineering. 2017; 5(1):4, https: / / doi.org / 10.3390 / jmse5010004).

[0023] The main cases where fluorescent dyes are not necessary are known in ophthalmology, where autofluorescence of ocular cells can provide valuable information about specific patient pathologies (see the article by A. Gakamsky et al. "Tryptophan and Non-Tryptophan Fluorescence of the Eye Lens Proteins Provides Diagnostics of Cataract at the Molecular Level", Scientific Reports 7, 40375 (2017), https: / / doi.org / 10.1038 / srep40375). In such cases, fluorescence microscopy can provide a rapid and reliable detection system using intrinsic fluorescence.

[0024] However, there is currently no optimized system to distinguish objects from bacteria at this level of accuracy, with proven application in microbiology. A.8 Hyperspectral imaging methods

[0025] Hyperspectral imaging solutions have been studied for the quantification of bacteria. These solutions generally use white light illumination in the VNIR (Visible and Near-infrared) spectrum, between 400nm and 1100nm, combined with machine learning (ML) or deep learning (DL) models.

[0026] These models are monobloc, i.e. feature extraction is performed to form the ML input, or a DL is used to process the hyperspectral images that may have been preprocessed. This makes it difficult to know which part of the images is associated with a bacterium and therefore limits the possibilities for validation during training, or requires imaging times that are incompatible with industrial implementation.

[0027] There are two main competing approaches: - one approach is based on measuring adulteration. For example, in the article by Zheng et al. "A Nondestructive Real-Time Detection Method of Total Viable Count in Pork by Hyperspectral Imaging Technique" , Applied Sciences 7, no. 3: 213, 2017, https: / / doi.org / 10.3390 / app7030213, a model to correlate the spectral signature of pork meat with the total viable cell count (TVC) is trained to rapidly quantify the TVC in a piece of pork. However, this measure mainly detects product adulteration rather than the actual amount of bacteria. It is effective for matrices spoiled by bacteria, but is not suitable for low bacterial loads (e.g., less than 100 CFU / mL) or matrices such as environmental samples and milk. - one approach is based on sorting bacteria based on their spectral signature on a microscope slide, as described in the article by Michael et al., "Hyperspectral imaging of common foodborne pathogens for rapid identification and differentiation", Food Science & Nutrition, 2019, 7: 2716-2725, https: / / doi.org / 10.1002 / fsn3.1131. This approach uses white light and transmittance microscopy to differentiate Salmonella spp, E. coli and Cronobacter sakazakii from each other. Although these hyperspectral solutions can differentiate between bacterial species, they require extensive sample preparation and high-resolution optics. In addition, they do not allow for the direct calculation of colony-forming units per milliliter (CFU / mL).

[0028] Beyond the problems mentioned above, hyperspectral imaging-based methods suffer from acquisition speed problems with existing hyperspectral cameras, which make them unusable in contexts of detecting low levels of bacteria. A.9 Multispectral imaging methods

[0029] Multispectral imaging solutions have been studied for the study of sample adulteration. Generally speaking, the methods for studying adulteration are unsuitable for the problem addressed by the invention.

[0030] As an example, the article by Spyrelli et al. "Implementation of Multispectral Imaging (MSI) for Microbiological Quality Assessment of Poultry Products", Microorganisms 2020, 8, 552. https: / / doi.org / 10.3390 / microorganisms8040552, describes the capture of reflectance at multiple wavelengths between 405nm and 970nm. These measurements are processed to form input vectors for a machine-learning process, which returns a TVC.

[0031] A similar approach is presented in the article by Fengou et al. "Detection of Meat Adulteration Using Spectroscopy-Based Sensors", Foods 2021, 10, 861. https: / / doi.org / 10.3390 / foods 10040861. In this article, the adulteration detection performance is validated against an ultraviolet fluorescence approach.

[0032] All these approaches have the disadvantage of not being able to detect low levels of bacteria. Furthermore, they are based solely on reflectance. By their nature, they are unsuitable for detecting single bacteria, both in terms of the resolution of the processed image and the signal used.

[0033] A.10 Excitation and emission matrix methods

[0034] Excitation and emission matrices (EEM) are used for multi-component analysis and provide a molecular fingerprint for different types of samples. Spectroscopic EEM has been used in different applications, particularly in the wine and oil industry, to predict taste by determining the specific compounds of interest (see the article by B. Quintanilla-Casas et al., "Using fluorescence excitation-emission matrices to predict bitterness and pungency of virgin olive oil: A feasibility study", Food Chemistry , Volume 395, 30 November 2022, 1336022022, https: / / doi.Org / 10.1016 / j.foodchem.2022.133602).

[0035] Compared to conventional fluorometry methods, EEM spectroscopy has the advantage of not being concentration dependent and of accurately identifying certain samples.

[0036] Some applications in microbiology have allowed, for example, the detection of bacteria in water at levels as low as 10 CFU / mL (see for example the article by Nakar et al., "Quantification of bacteria in water using PLS analysis of emission spectra of fluorescence and excitation-emission matrices", Water Research, Volume 169, 2020, 115197, https: / / doi.Org / 10.1016 / j.watres.2019.115197). However, in the case of food, these methods are not capable of detecting anything other than adulteration, and in particular TVCs below 100 CFU / mL.

[0037] Finally, they require the calibration of tryptophan to obtain accurate results and the use of expensive equipment such as a fluorimeter. B. Summary of the Invention

[0038] Generally speaking, the Petri dish method and the exploratory methods described above are not satisfactory for one or more of the following reasons: implementation time, complexity, cost, detection quality, detection threshold, ability to determine the TVC.

[0039] The invention improves the situation. To this end, it proposes, according to a first aspect, a device for detecting bacteria in a sample which comprises an optical bench comprising an illumination block comprising an optical source and a distribution optic for illuminating a sample, and a measurement block comprising a collection optic and a measurement optic, the collection optic being arranged to redirect measurement radiation coming from the sample in response to illumination towards the measurement optic. The optical source is arranged to emit a first radiation of wavelength substantially equal to 365nm, and a second radiation of wavelengths in a range between 375nm and 1000nm to the distribution optics in order to illuminate the sample, the distribution optics comprises a magnification objective, and the collection optics further comprises one or more splitters arranged to separate the measurement radiation into at least a first beam and a second beam, the first beam having a portion of the measurement radiation having one or more wavelengths in a first range between 400nm and 414nm, the second beam having a portion of the measurement radiation having one or more wavelengths in a second range between 414nm and 490nm.The measuring optics comprises at least a first optical sensor for the first beam, and a second optical sensor for the second beam arranged to each return an image in which each pixel is associated with an area of the sample illuminated by the magnification objective and with a measured light amplitude value respectively.This device is arranged to control a relative movement of the sample with respect to the illumination block in order to illuminate successive portions of the sample with the first radiation and with the second radiation and to derive therefrom a set of multispectral images for each of the successive portions of the sample, and further comprises an analyzer comprising a classifier and a calculator, the classifier comprising a deep learning neural network arranged to receive the set of multispectral images associated with a portion of the sample and to return an image in which each pixel is associated with an element type identifier chosen from a list comprising at least two elements, one of which designates a bacterium, and the calculator being arranged to determine bacteria detection data from the images returned by the deep learning neural network.

[0040] This device is particularly advantageous because it provides a device that can carry out bacteria detection operations at very low detection levels, and in times compatible with an industrial production line.

[0041] According to various embodiments, this device may have one or more of the following characteristics: - said one or more separators of the collection optics is further arranged to separate the measurement radiation into a third beam distinct from the first beam and the second beam, this third beam having a portion of the measurement radiation having one or more wavelengths in a third range between 470nm and 1000nm, the measurement optics comprising a third optical sensor (44) for measuring the third beam, the analyzer being further arranged to receive the measurements from the third optical sensor with the measurements from the first optical sensor and the second optical sensor to extract bacteria detection data, - the optical bench further includes an autofocus arranged between the optical source and the magnification objective, - the magnification lens and autofocus are included in the distribution optics and in the collection optics, - the deep learning neural network of the classifier is a convolutional neural network, - the deep learning neural network of the classifier is a transformer, - the classifier comprises at least one convolutional neural network coupled to a transformer, - the calculator is arranged to determine a ratio between the surface area occupied by pixels associated with a bacteria identifier and a reference surface area, and to calculate bacteria detection data from this ratio, - the classifier is arranged to associate a pixel with an identifier with an element type identifier chosen from a list comprising a bacterium, a matrix element, an air bubble, a membrane element or a foreign element, and - the calculator is arranged to determine the reference surface from the pixels whose identifiers are associated with a membrane element.

[0042] According to a second aspect, the invention also relates to a method for detecting bacteria in a sample comprising the following operations: a) determining one or more sets of multispectral images, each set of multispectral images of a given sample portion comprising images obtained by measuring, in a first wavelength range between 400nm and 414nm and in a second range between 414nm and 490nm, an autofluorescence beam emitted by the given sample portion when it is illuminated by radiation of wavelength substantially equal to 365nm, and on the other hand a reflectance beam of the given sample portion when it is illuminated by radiation of wavelengths in a range between 375nm and 1000nm,b) providing each set of multispectral images to a deep learning neural network to produce an image in which each pixel is associated with an element type identifier chosen from a list comprising at least two elements, one of which designates a bacterium, and c) calculating bacteria detection data from at least some of the images of operation b).,

[0043] According to various embodiments, this method may have one or more of the following characteristics: - operation a) comprises determining a set of multispectral images for a set of non-overlapping sample portions whose union covers the entire surface of the sample, - operation a) is performed for a portion of sample, then operation b) is performed for this portion of sample, and operations a) and b) are repeated until an end-of-measurement condition including exceeding a threshold or determining a set of multispectral images for the entire sample is encountered, - operation a) is performed for a chosen number of sample portions, then a necessary number of sample portions is determined by applying operation b) on the resulting multispectral image sets, then operations a) and b) are performed with a stopping condition taking into account the necessary number of sample portions, - the number of sample portions required is re-evaluated each time operations a) and b) are carried out, and in which the stopping condition takes into account the re-evaluated number of sample portions required, and - operation c) comprises determining a ratio between the area occupied by pixels associated with a bacteria identifier and a reference area, and calculating bacteria detection data from this ratio.

[0044] Other characteristics and advantages of the invention will appear more clearly on reading the following description, taken from examples given for illustrative and non-limiting purposes, taken from the drawings in which: - [Fig.l] represents a schematic view of a device according to the invention, - [Fig.2] represents an embodiment of the optical bench of [Fig.l], - [Fig.3] represents an example of an image obtained at the output of the optical block of [Fig.2], - [Fig.4] represents a generic view of an analyzer implemented in the embodiment of [Fig.l], - [Fig.5] represents a block diagram of an implementation of the analyzer of [Fig.4], - [Fig.6] represents an example of an image obtained as output from the classifier of [Fig.5], - [Fig.7] represents a diagram illustrating a method of operation of the analyzer of [Fig.4] according to an embodiment in operational mode, - [Fig.8] represents the results of detection of bacteria obtained with the matrix at the origin of figures 3 and 6, - [Fig.9] represents images similar to those of figures 3 and 6 obtained with another matrix, - [Fig.10] represents bacteria detection results obtained with the matrix at the origin of [Fig.9], and

[0045] - [Fig.l 1] represents bacteria detection results obtained with an analyzer using a transformer.

[0046] The drawings and the description below contain, for the most part, elements of a certain character. They may therefore not only serve to better understand the present invention, but also contribute to its definition, if necessary. C. Examples of embodiments of the invention

[0047] [Fig.l] represents a schematic view of an embodiment of a device 2 according to the invention.

[0048] The device 2 comprises an optical bench 4 and an analyzer 6. The optical bench 4 comprises an illumination block 8 and a measuring block 10.

[0049] The optical bench 4 is arranged to carry out multispectral imaging operations on a sample 12 obtained by grinding and depositing on a membrane a sample 14 for which a bacteria detection value 16 is sought. In the embodiment described here, the sample may be wheat, flour or even a liquid reflux (for example to test the presence of bacteria in water). In the following, the sample 14 may also be called a "matrix".

[0050] At the output of the optical bench 4, a plurality of multispectral images 18 (hereinafter referred to as a set of multispectral images) are transmitted to the analyzer 6 to determine bacteria detection data 16 which may be a bacteria quantity value. In some embodiments, the bacteria quantity value may be a number of colony forming units per milliliter (CFU / mL). In other embodiments, this bacteria quantity value may be a Boolean value indicating whether a threshold of colony forming units per milliliter (CFU / mL) is crossed. Although reference is made here to colony forming units per milliliter, any other relevant TVC measurement may be retained, such as colony forming units per gram, colony forming units per square centimeter, etc.As will be seen below, in certain embodiments, the bacteria detection data 16 may be data making it possible to determine a bacteria quantity value by processing or normalization.

[0051] In the embodiment of [Fig.l], the analyzer 6 implements a machine learning model (ML). For this reason, the broken arrows are used to show the elements for carrying out the learning of the analyzer 6, via a Petri dish 20 in which the bacteria present in the sample 14 are conventionally cultured, then counted to provide data for the learning of the analyzer 6. Once the learning has been carried out, only the solid arrows are implemented by the device 2 to calculate a bacteria detection value 16. The operation of the analyzer 6, its drive and its variant embodiments will be described further with the embodiments of figures 4 to 7. Cl Optical Bench

[0052] [Fig.2] illustrates an embodiment of the optical bench 4.

[0053] The optical bench 4 comprises an illumination block 22 and a measuring block 24.

[0054] The illumination unit comprises an optical source 26 and a distribution optic 28 for illuminating the sample 12. In the example described here, the distribution optic 28 comprises a magnification objective 30 and an autofocus 32. As will be seen below, the magnification objective 30 and the autofocus 32 also serve as collection optics for carrying out the measurements. In certain embodiments, the autofocus 32 may be omitted.

[0055] In the example described here, the optical source 26 is a white light source. Generally, the optical source 26 may have an emission spectrum comprising wavelengths between 360 nm and 1000 nm. The reason for this choice is that the Applicant has discovered that it is possible, using a multispectral approach, to study the spectrum in autofluorescence (also called "intrinsic fluorescence") and in reflectance in order to distinguish the presence of bacteria in matrices such as water, flour, wheat or other products to be tested.

[0056] This is highly unusual. Indeed, while reflectance is an interesting signal in the context of adulteration detection at the macroscopic level, it is a priori irrelevant when detection levels are much lower, for example at the level of each bacterium taken individually. Traditionally, reflectance is ignored in microscopy in favor of the transmittance of samples by phase contrast imaging. However, the study of transmittance by phase contrast imaging is impossible in the present case.

[0057] Against all expectations, the Applicant discovered that not only do bacteria exhibit a useful reflectance signal, but also, unlike all existing approaches, the combination of autofluorescence and reflectance makes it possible to detect bacteria individually and to exclude foreign bodies which could distort the measurements.

[0058] The invention therefore combines the non-conventional use of reflectance with autofluorescence, where the only known methods - which do not otherwise address the problem targeted by the invention - use at best a single type of signal.

[0059] In the example described here, the optical source 26 is a light-emitting diode (LED) source. Alternatively, any optical source capable of emitting radiation capable of causing autofluorescence in the range mentioned above will be suitable provided that the distribution optics 28 can concentrate the radiation on the sample 12 to obtain a suitable size for the shots. Thus, the optical source 26 could be a laser source or any other suitable light source.

[0060] Since autofluorescence is a low-intensity phenomenon, the optical source 26 is arranged to emit autofluorescence radiation with a wavelength substantially equal to 365 nm, i.e. centered on the wavelength 365 nm, which corresponds to the autofluorescence excitation of NADH (which is the reduced form of nicotinamide adenine dinucleotide). Thus, the illumination radiation does not interfere with the measurement of autofluorescence. As regards reflectance, the situation is different. Indeed, reflectance can be measured over a broad spectrum, from 300 nm to more than 1000 nm. However, in order to limit the risk of photobleaching, the optical source is arranged to emit reflectance radiation with a wavelength between 375 nm and 1000 nm. Preferably, the reflectance radiation may have a wavelength of between 400nm and 1000nm.

[0061] Furthermore, the multispectral approach is advantageous over a hyperspectral approach because it allows for faster imaging throughput for a given resolution. More specifically, the multispectral approach, with a faster acquisition throughput, allows for higher resolution images to be produced with a constant time budget, and therefore for the detection of colony forming unit (CFU / mL) levels. Thus, the 30x magnification objective in the example described here is a 20x magnification objective, which allows for images of approximately 0.5mm x 0.5mm to be taken.

[0062] Alternatively, other objectives could be used. Generally speaking, the combination of the objective and the cameras used results in a resolution of around 84 pixels for a bacterium. Generally speaking, any combination that achieves a resolution of between 1 pixel and 1500 pixels (which represents a magnification of approximately 100X) makes it possible to obtain detection whose quality and imaging duration remain compatible with the desired performance, i.e. the ability to perform a reliable measurement in less than an hour.

[0063] The sample 12 is received on a movable plate 33 of the optical block 4. The movable plate 33 makes it possible to move the sample 12 in an XY plane relative to the measuring block 22 and to the measuring block 24, which makes it possible to illuminate and measure a different portion of the sample 12 each time. The autofocus 32 can also be controlled as a function of the movement of the plate 33 in order to carry out the focus adjustment in Z. As will be seen below, the movement by the movable plate 33 can be controlled by the analyzer 6. Alternatively, the movable plate 33 can be controlled by a control module of the optical bench 4. Still as a variant, it is the objective 30 which could be movable relative to the sample 12.

[0064] The measuring block 6 comprises a collection optic 34 and a measuring optic 36. The role of the collection optic 34 is to recover the autofluorescence and reflectance radiation from the sample 12 and to redirect it towards the measuring optic 36.

[0065] As mentioned above, the optical bench 4 is designed compactly, and the magnification objective 30 and the autofocus 32 serve both the distribution optics 28 and the collection optics 34. Thus, the light beam from the light source 26 is focused and its size is reduced on the sample 12 by the objective 30 and the autofocus 32, and the radiation which comes from the sample 12 passes back through the objective 30 and the autofocus 32 which restores its initial size for the optical measurement.

[0066] In the example described here, the collection optics 34 further comprises three separators 38 and three lenses 39 which each redirect a portion of the radiation coming from the sample 12 towards the measurement optics 36.

[0067] Thus, a first separator 38 comprises a dichroic mirror which isolates a portion of the radiation from the sample 12 whose wavelengths are between 400nm and 425nm, and preferably between 400nm and 414nm to form a first beam. This dichroic mirror directs the remainder of the radiation from the sample 12 towards a second separator 38.

[0068] Similarly, the second separator 38 comprises a dichroic mirror 38 which isolates a portion of the radiation from the sample 12 whose wavelengths are between 414nm and 490nm, and preferably between 414nm and 470nm to form a second beam. This dichroic mirror directs the remainder of the radiation from the sample 12 towards a third separator 38.

[0069] Finally, the third separator 38 recovers the remainder of the radiation which therefore has wavelengths beyond 470nm to form a third beam. Alternatively, the third separator 38 may comprise a dichroic mirror which selects a portion of the radiation from the sample 12 whose wavelengths are preferably between 470nm and 520nm to form the third beam.

[0070] The first beam, the second beam and the third beam are each received by a respective lens 39 of the collection optics 34.

[0071] The measuring optics 36 comprises in the example described here a first camera 40, a second camera 42 and a third camera 44. Each lens 39 respectively redirects the first beam towards the first camera 40, the second beam towards the second camera 42 and the third beam towards the third camera 44.

[0072] The first camera 40, the second camera 42 and the third camera 44 are arranged to measure spectral bands which correspond to the wavelength range of respectively the first beam, the second beam and the third beam.

[0073] In some embodiments, the third separator 38 may be omitted, and the analyzer 6 will operate on the measurements of the first beam and the second beam only. In this case, as will be seen below, instead of two image triplets, the output of the cameras 40 to 44 will then be two pairs of images (one pair for the measurement of autofluorescence and another pair for the measurement of reflectance).

[0074] In the example described here, the first camera 40, the second camera 42 and the third camera 44 are of the Phoenix 8.1 MP Model (IMX566), and the image resulting from each shot has a resolution of 1420*1420, as shown in [Fig.3]. Other models of cameras or optical sensors could of course be used, as well as other resolutions for the corresponding images. It will nevertheless be necessary to take into account the fact that the diffraction limit sets a maximum useful resolution.

[0075] In each image, each pixel corresponds to an area of the sample 12 that was illuminated. The value associated with each pixel corresponds to the sum of the intensities measured for the wavelengths of the camera considered at this location. To represent the value associated with each pixel, a color can be used that depends on the sum of the measured intensities.

[0076] Thus, in operation, the optical source 26 will successively illuminate portions of the sample 12. The illumination of each portion of the sample 12 is carried out in two stages.

[0077] A first illumination of approximately 1 second at a wavelength of 365nm is used to elicit the autofluorescence radiation from the sample 12 which is measured in three beams by the three cameras 40 to 44, from which three images of the sample portion similar to that of [Fig.3] emerge. Then, a second illumination of approximately 50ms is carried out at approximately 405nm, and the reflectance radiation of the sample 12 is measured in three beams by the three cameras 40 to 44, from which three other images of the sample portion similar to that of [Fig.3] emerge.

[0078] It is these sextuplets of images (two times three images respectively in autofluorescence and in reflectance) which are processed by the analyzer 6 in order to determine bacteria detection data 16 making it possible to return a bacteria detection value. As mentioned above, in the case where only two bands are measured, the analyzer 6 will be arranged to process quadruplets of images.

[0079] In the example described here, autofluorescence and reflectance are measured by the same cameras because the optical source 26 includes a filter for producing the autofluorescence radiation which is centered at 365nm and is slightly permeable to radiation around 405nm. Thus, the reflectance measurement can be performed by the cameras 40 to 44 by greatly increasing the amplitude of the optical source 26 for the reflectance measurement.

[0080] Alternatively, the optical source 26 could comprise a mechanism alternating a 365nm filter with a 405nm filter between the autofluorescence measurement and the reflectance measurement. Still as a variant, off-axis illumination in white light with a wavelength greater than 400nm could be used to carry out the reflectance measurement. In all cases, two image triplets (or two pairs of images) are always recovered, on the one hand for the autofluorescence and on the other hand for the reflectance. These two image triplets (or two pairs of images) will subsequently be called “image sextuplets” (“image quadruplets” in the case of two bands, and therefore two pairs), or even “multispectral image set”. C.2 Analyzer

[0081] The analyzer 6 and its processing of the measurements from the cameras 40 to 44 will now be described with reference to [Fig.4].

[0082] The analyzer 6 comprises a data storage 100 and a calculator 102.

[0083] The data storage 100 receives sample image data, deep neural network data, and bacteria detection data.

[0084] The sample image data are each associated with coordinates making it possible to know from which part of the sample 12 they come, as well as camera data, making it possible to know which of the first beam, second beam or third beam they are associated. As explained above, the sample image data comprise pixels with each of which is associated an intensity value representative of the sum of the intensities measured by the camera for the location of the sample 12 designated by the coordinates and the pixel concerned. In addition, an additional camera could be used to detect the absolute position of the sample 12. Still alternatively, the sample rack 12 can be adapted for the same purposes.

[0085] Data storage 100 may be any type of data storage suitable for receiving digital data: hard disk, flash memory hard disk, flash memory in any form, RAM, optical disk, locally or cloud distributed storage, etc.

[0086] The computer (PC) 102 may include one or more processors (P) 104, a network interface 106 comprising a transmitter (Tx) 108 and a receiver (Rx) 110 to enable the apparatus to transmit and / or receive data to other computing devices connected to a network 112 (e.g., an IP network) to which the network interface 106 is connected.

[0087] In some embodiments in which the computer 102 includes a programmable processor, a computer program product (CPP) 116 may be provided to implement the phase correction processing. The computer program product (CPP) 116 stores a computer program (CP) 118 that includes computer readable instructions (CRI) 120. The computer program product (CPP) 116 may be stored on a computer readable medium (CRM) 122, which may be a non-transitory computer readable medium, such as magnetic media (hard drive, SSD, magnetic tape, etc.), optical media (CD, DVD, Blu-ray, etc.), memory (RAM, flash, etc.), distributed or cloud storage, etc.The computer readable medium (CRM) 122 may also be stored in the data storage 100.

[0088] The processor(s) P 104 may be any processor suitable for the calculations described below. Such a processor may be implemented in any known manner, in the form of a microprocessor for a personal computer, laptop, tablet or smartphone, a processor dedicated to signal processing ("DSP" in English), a dedicated chip of the FPGA or SoC type, a computing resource on a grid or in the cloud, a cluster of graphics processing units ("GPUs"), a microcontroller, or any other form suitable for providing the computing power necessary for the implementation described below. One or more of these elements may also be implemented in the form of specialized electronic circuits such as an ASIC. A combination of processor(s) and electronic circuits may also be considered. Processors dedicated to machine learning may also be considered.

[0089] In the exploratory methods described in the introduction, particularly in some of the multispectral or hyperspectral methods, a deep learning neural network is used to directly extract the value of colony forming units per milliliter (CFU / mL), from a single shot of the sample 12.

[0090] Applicant explored several ways to process the sample image data and determined that using a deep-learning neural network on a single image is not desirable.

[0091] On the contrary, the Applicant has discovered that, in order to be able to calculate an industrially usable bacteria detection value, i.e. in orders of magnitude of 1 CFU / mL, it is desirable to have a spatial resolution of high data collection, which requires taking more than 100 shots for a 500 mm sample size 2 approximately. Advantageously, concentrating the light from the optical source also makes it possible to increase the surface intensity of illumination and therefore the corresponding autofluorescence response.

[0092] It should first be noted that this number of shots makes the use of hyperspectral cameras impossible because the acquisition time for each shot is 2.5 minutes. But beyond this, the Applicant realized that the lower the level of colony-forming units per milliliter (CFU / mL) sought, the less reliable the direct detection by deep learning neural network. Indeed, with the increase in resolution, the information on the presence of bacteria becomes diluted, or even spread over several images, and it is difficult for the neural network to learn reliably.

[0093] The Applicant has thus discovered that it is preferable to carry out the detection in two operations.

[0094] For this reason, as shown in [Fig.5], the analyzer 6 comprises a classifier 50 which comprises a deep learning neural network which receives a set of multispectral images as input and returns an image 52 as output, and a calculator 54 which receives the image 52 as input and returns the detection data 16.

[0095] It goes without saying that the distinction between the classifier 50 and the calculator 54 is essentially functional, but that the analyzer 6 can be seen as made of a single block.

[0096] Thus, a first detection operation implemented by the classifier 50 consists of associating, from all the images concerning a given area of sample 12, a label with each pixel of this area. Indeed, the sample 12 is typically composed of a membrane and the atomized sample 14. Consequently, the optical measurement can therefore, for each pixel, correspond to the following natures: membrane material, matrix element, a bacterium, an air bubble or a foreign element such as dust. In its simplest version, the label can be reduced to “bacteria” or “other element”.

[0097] In other words, each pixel receives a value that indicates whether this pixel is considered to belong to the membrane of the sample 12, to a bacterium, to a matrix element that is the subject of the bacteria detection measurement, to an air bubble, or to a foreign element.

[0098] Alternatively, the deep learning neural network may return a vector that indicates a probability for each pixel to belong to a given category (e.g., [0.4; 0.6] would indicate that there is a 40% chance that a given pixel belongs to a bacterium and a 60% chance that it does not in the case of a binary label described above). In this case, the first module 50 may be arranged to determine a unique label for each pixel, either by choosing the label with the highest value, or by using in part the value of the neighbors, etc. Alternatively, the probability of belonging to the “bacteria” category could be taken into account in the subsequent calculation of the bacteria detection data.

[0099] [Fig.6] represents an example of image 52 obtained after applying the labels based on the measurements shown in [Fig.3].

[0100] The Applicant tested several deep learning neural networks for this task, and found that convolutional neural networks (CNNs), recurrent neural networks (RNNs) and transformers or self-attentive models (Transformers) are useful for performing this first operation.

[0101] To train the deep neural network, multispectral image sets were measured for many samples of each matrix, and each pixel in each sextuplet of images was annotated to indicate its corresponding label. By each pixel in each sextuplet it should be understood that a label is associated with a pixel coordinate for six images (both triplets) corresponding to the same shot.

[0102] The Applicant has tested several of these CNN models including Unet, Unet++, MAnet, Linknet, FPN, PSPNet, PAN, DeepLabV3 and DeepLabV3+. The address https: / / web.archive.org / web / 20231021175139 / https: / / smp.readthedocs.io / cn / latcst / provides access to these CNNs. Among the transformers, SegViTv2 (see https: / / arxiv.org / pdf / 2306.06289.pdf), Segmenter (see , SegFormer (see https: / / huggingface.co / docs / transformers / model_doc / segformer) have also been identified. The Applicant has also determined that a combination of CNN, RNN and / or transformer can be useful.

[0103] A second part of detection is carried out by the computer 54 and consists of quantifying the presence of bacteria on the basis of the images 52 on which the pixels have received a label. This can be carried out in various ways.

[0104] According to one embodiment of the invention, the surface area of pixels associated with a bacterium is calculated and accumulated image by image to obtain a total surface area occupied by the bacteria on the sample 12, which makes it possible to detect a TVC value by reducing it to the detected membrane surface area. To determine the TVC from the ratio between the surface area occupied by the bacteria and the surface area occupied by the membrane, Petri dish cultures were carried out to measure the actual TVC for these samples, from standardized quantities of sampling 14.

[0105] The determined TVC value obviously depends on the quantity of sample 14 which was used to make the sample 12. Consequently, if the quantity of sample 14 is not standardized, then the analyzer 6 will be able to receive as input the value in grams, liters or square centimeters allowing the TVC value to be reduced to CFU / g, CFU / mL or CFU / cm 2Alternatively, this standardization may be carried out later.

[0106] By "frame by frame" is meant here the images obtained after processing by the first operation above, and which therefore each correspond to a given portion of the sample which has been illuminated and whose autofluorescence and reflectance response has been measured. As suggested above, the cumulative TVC value can also be compared to a threshold, and the operation can stop as soon as the threshold has been crossed or the entire sample 12 has been illuminated.

[0107] This type of processing is illustrated in [Fig.7], which shows an operating loop of device 2.

[0108] This loop begins with an operation 60 in which a set of multispectral images of a portion of the sample 12 is obtained by imaging by the optical bench 4.

[0109] Then, in an operation 62, the analyzer 6 calculates a ratio of bacteria surface area to membrane surface area, and a bacteria detection value is calculated in an operation 64.

[0110] In operation 64, a bacteria detection value may be obtained by determining a sample portion TVC for the set of multispectral images of operation 60, this sample portion TVC may then be accumulated over all sample portions. Alternatively, the bacteria detection value may be re-evaluated at each sample portion, i.e., the bacteria area to membrane area ratio for the current measurement is updated at each iteration of the loop, and the TVC is evaluated from this total ratio (up to the current loop).

[0111] Then, in an operation 66, it is determined whether the loop should continue. In the case of detection of a threshold, this involves comparing the result of operation 64 with this threshold. Otherwise, this operation consists of determining whether there is another portion of sample left to illuminate. Finally, the loop ends in an operation 68 with the sending of the bacteria detection data thus determined.

[0112] [Fig.8] shows the performance of the measurement made by the loop of the [Fig.7] method.

[0113] [Fig.9] shows respectively the images corresponding to [Fig.3] (left) and [Fig.6] (right) obtained with wheat as matrix, and [Fig.10] shows the performance of the measurement carried out by this method with wheat as matrix.

[0114] The Applicant has further discovered that it is possible to speed up the measurement by reducing the number of shots taken.

[0115] More specifically, in a particular embodiment, 5 (for example) non-overlapping images are obtained to initialize the measurement loop. Based on these first values, a number of shots allowing the calculation of the TVC value with a precision of ±0.5 LOG.

[0116] The table below represents numbers of shots based on TVC values.

[0117] In a first variant, a loop similar to that of [Fig.7] can be implemented, but with a fixed number of shots taken from the table above as a stopping condition.

[0118] In a second variant, the loop can be slightly modified by evaluating the number of shots at each loop iteration based on the evolution of the TVC value of the previous loop, and by defining a loop end condition based on the comparison between the updated number of shots required and the number of loops already performed and / or the exceeding of a TVC value threshold if applicable.

[0119] The Applicant's work suggests that there is no preferred route for taking photographs. However, in order to limit any risk, it is preferable to go through the sample by choosing photographs that are roughly equally distributed among themselves. For example, it is possible to define non-overlapping areas, and to take a measurement in each area, then repeat.

[0120] In other variations, some of the 52 images could be skipped.

[0121] Thus, the device described above makes it possible to carry out a TVC measurement with great precision in very low detection ranges, with a measurement duration which is less than 2 hours, and is regularly less than 10 minutes.

[0122] As mentioned above, the Applicant has also experimented with the use of transformers as deep learning neural networks. Indeed, despite their qualities demonstrated above, CNNs have narrow effective receptive fields.

[0123] Transformers introduce a non-local self-attention mechanism that can perform operations combining distant pixels in the image. This significantly increases the effective receptive field and improves the model's performance for a given computing power.

[0124] To validate this theory, different machine learning models were tested on a dataset consisting of 56 images (10 for validation and 46 for training). In this dataset, 11 samples contained only wheat and 45 samples contained only bacteria. These samples were annotated at the pixel level.

[0125] The performance of transformer models of different sizes and those of a convolutional neural network model. The transformer used is of the Segformer type (bO, bl, b2, b3 - bO being the one requiring the least computing power and b3 the most), while the convolutional neural networks used are DeeplabV3+ coupled with Resnet 101. In the latter case, the DeeplabV3+ neural network is used as a "backbone” to extract a set of features that is processed by the Resnet neural network used as a "head” to determine the TVC value. The backbone transforms the image into features that are then exploited by the head which will, in classic applications, either perform classification, regression, object detection, or in the example described here, semantic segmentation (one class per pixel).

[0126] It is also possible to combine a convolutional or recurrent neural network together with a transformer. In this case, the transformer will always be used as the "head”, while the other neural network will be used as the "backbone”.

[0127] The models are trained on an Nvidia T4 GPU with 4 Intel Cascade Lake vCPUs and 16 GB of RAM.

[0128] This comparison has two parts:

[0129] 1. the intersection over the union (IoU or “Intersection over Union” in English) on the validation set (the 10 images mentioned above)

[0130] 2. end-to-end prediction of the total viable bacteria count (TVC) on 5 samples not used during training, all composed of 5 images that mix wheat and bacteria (remembering that the model was only trained on pure samples). The total number of viable bacteria obtained is then compared with that measured in the laboratory using Petri dishes in terms of absolute error and precision. Precision is defined as the number of samples for which the absolute error is less than 0.5.

[0131] The table below summarizes the results of the evaluations of the first part:

[0132] The results show that using transformation models has a clear advantage.

[0133] - All transformer models outperform the CNN model in terms of intersection over union (loU) on the validation set.

[0134] - All transformers are faster than CNN in terms of number of iterations per minute.

[0135] - SegFormer b2 and b3 outperform CNN in terms of average TVC error.

[0136] - SegFormer b3 matches CNN in terms of TVC accuracy.

[0137] [Fig. 11] shows the results of the second part of the evaluations. This figure shows in particular the results obtained using the SegFormer b3 transformer, as well as their comparison with the Petri dish measurements.

Claims

Claims

1. Device for detecting bacteria in a sample, comprising an optical bench (4) comprising an illumination block (22) comprising an optical source (26) and a distribution optic (28) for illuminating a sample, and a measurement block (24) comprising a collection optic (34) and a measurement optic (36), the collection optic (34) being arranged to redirect measurement radiation coming from the sample (12) in response to illumination towards the measurement optic (36), characterized in that the optical source (26) is arranged to emit a first radiation of wavelength substantially equal to 365nm, and a second radiation of wavelengths in a range between 375nm and 1000nm to the distribution optic (28) in order to illuminate the sample (12), the distribution optic (28) comprises a magnification objective (30),the collection optics (34) further comprises one or more splitters (38) arranged to separate the measurement radiation into at least a first beam and a second beam, the first beam having a portion of the measurement radiation having one or more wavelengths in a first range between 400nm and 414nm, the second beam having a portion of the measurement radiation having one or more wavelengths in a second range between 414nm and 490nm, the measurement optics (36) comprising at least a first optical sensor (40) for the first beam, and a second optical sensor (42) for the second beam each arranged to return an image in which each pixel is associated with an area of the sample (12) illuminated by the magnification objective (30) and with a measured light amplitude value respectively,the device (2) being arranged to control a relative movement of the sample (12) with respect to the illumination block (22) in order to illuminate successive portions of the sample (12) with the first radiation and with the second radiation and to obtain therefrom a set of multispectral images for each of the successive portions of the sample (12), the device (2) further comprising an analyzer (6) comprising a classifier (50) and a calculator (54), the classifier (50) comprising a deep learning neural network arranged to receive the set of multispectral images associated with a portion of the sample (12) and to return an image (52) in which each pixel is associated with an element type identifier chosen from a list comprising at least two elements, one of which designates a bacterium, and the calculator (54) being arranged to determine bacteria detection data (16) from the images returned by the deep learning neural network.

2. Device according to claim 1, wherein said one or more splitters (38) of the collection optics (34) is further arranged to separate the measurement radiation into a third beam distinct from the first beam and the second beam, this third beam having a portion of the measurement radiation having one or more wavelengths in a third range between 470nm and 1000nm, the measurement optics (36) comprising a third optical sensor (44) for measuring the third beam, the analyzer (6) being further arranged to receive the measurements from the third optical sensor (44) with the measurements from the first optical sensor (40) and the second optical sensor (42) to derive the bacteria detection data (16) therefrom.

3. Device according to claim 1 or 2, wherein the optical bench (4) further comprises an autofocus (32) disposed between the optical source (26) and the magnification objective (30).

4. Device according to claim 3, wherein the magnification objective (30) and the autofocus (32) are included in the distribution optics (28) and in the collection optics (34).

5. Device according to one of the preceding claims, in which the deep learning neural network of the classifier (50) is a convolutional neural network.

6. Device according to one of claims 1 to 4, in which the deep learning neural network of the classifier (50) is a transformer.

7. Device according to one of claims 1 to 4, in which the classifier (50) comprises at least one convolutional neural network coupled to a transformer.

8. Device according to one of the preceding claims, in which the calculator (54) is arranged to determine a ratio between the surface occupied by pixels associated with a bacteria identifier and a reference surface, and to calculate the bacteria detection data (16) from this ratio.

9. Device according to one of the preceding claims, in which the classifier (50) is arranged to associate a pixel with an identifier with an element type identifier chosen from a list comprising a bacterium, a matrix element, an air bubble, a membrane element or a foreign element.

10. Device according to claim 8 in combination with claim 9, in which the calculator (54) is arranged to determine the reference surface from the pixels whose identifiers are associated with a membrane element.

11. A method for detecting bacteria in a sample comprising the following operations: a) determining one or more sets of multispectral images, each set of multispectral images of a given sample portion comprising images obtained by measuring, in a first wavelength range between 400nm and 414nm and in a second range between 414nm and 490nm, an autofluorescence beam emitted by the given sample portion when it is illuminated by radiation of wavelength substantially equal to 365nm, and on the other hand a reflectance beam of the given sample portion when it is illuminated by radiation of wavelengths in a range between 375nm and 1000nm,b) providing each set of multispectral images to a deep learning neural network to produce an image (52) in which each pixel is associated with an element type identifier chosen from a list comprising at least two elements, one of which designates a bacterium, and c) calculating bacteria detection data from at least some of the images of operation b).,

12. The method of claim 11, wherein step a) comprises determining a set of multispectral images for a set of non-overlapping sample portions whose union covers the entire surface of the sample.

13. The method of claim 11, wherein operation a) is performed for a sample portion, then operation b) is performed for that sample portion, and operations a) and b) are repeated until a measurement end condition comprising exceeding a threshold or determining a set of multispectral images for the entire sample is encountered.

14. A method according to claim 11, wherein operation a) is performed for a selected number of sample portions, then a required number of sample portions is determined by applying operation b) to the resulting multispectral image sets, then operations a) and b) are performed with a stopping condition taking into account the required number of sample portions.

15. A method according to claim 14, wherein the number of sample portions required is re-evaluated each time operations a) and b) are performed, and wherein the stopping condition takes into account the re-evaluated number of sample portions required.

16. Method according to one of claims 11 to 15, in which operation c) comprises determining a ratio between the surface occupied by pixels associated with a bacteria identifier and a reference surface, and calculating the bacteria detection data (16) from this ratio.

Images