Devices for characterizing and comparing erythema areas

Abstract

The present invention relates to a system for comparing erythema areas, the system comprising: a multispectral camera (701) for acquiring multispectral images (711, 712) of each comparison area; a computer processing device (702) comprising storage, display, input and calculation devices, which is capable of retrieving images acquired by the multispectral camera, characterized in that the computer processing device comprises: an image processing program, which is capable of performing spectral unmixing of the acquired images and isolating the erythema component of the images; and a program for comparing the images based on the intensity of the erythema component of each image.

Description

Technical Field

[0001] The present invention relates to a device for characterizing and comparing erythema areas of the skin. The present invention is more particularly, but not exclusively, suitable for determining the sun protection index or sun protection factor (SPF) of photoprotective products intended to be applied to the skin. Background Art

[0002] The development and evaluation of cosmetics requires many tests, among which skin irritation tests are intended, for example, to check the safety of a product or its protection index.

[0003] These tests are performed by exposing the skin, human or animal, to the product or applying a stress thereto (such as by exposure to UV radiation, according to various means) and then detecting and characterizing the erythema area on this exposed area.

[0004] Characterization of the erythematous area is based on a comparison between the erythematous area and healthy skin (without erythema) or between two erythematous areas subjected to different conditions.

[0005] The detection of erythema and its characterization is based on the interpretation of the erythema condition by the practitioner (e.g., dermatologist).

[0006] Determining the sun protection factor of a photoprotective product is an example of a test that involves the detection and characterization of erythema areas.

[0007] Photoprotective products are products that are intended to come into contact with human skin and contain any component capable of absorbing, reflecting or scattering UV rays. Such products work by reflecting, absorbing or scattering radiation outside the skin and have no therapeutic effect on the skin.

[0008] Alternatively, the light protection product is a textile.

[0009] The SPF determines how long the product will increase the skin's UV exposure before it produces erythema (more commonly designated as "sunburn").

[0010] Thus, if a subject who exposes himself to a given UV radiation without protection develops erythema after 5 minutes of exposure to this radiation, the same subject who exposes himself to this same radiation using a photoprotective product with an SPF of 10 develops erythema after 50 minutes.

[0011] Determining the sun protection factor (SPF) of such products involves in vivo testing that involves exposing the protected and unprotected skin of test subjects to UV radiation in a controlled manner.

[0012] Therefore, the determination of the sun protection factor of products applied on the skin is defined by the ISO 24444:2010 standard, which consists of in vivo tests performed on the skin of multiple test subjects.

[0013] According to this test, for a given subject, a first, unprotected part of the skin (ie, to which no photoprotective product has been applied) is exposed to UV radiation according to increasing doses according to substantially circular spots.

[0014] A second portion adjacent to the first portion is covered by the protective product and is also exposed to UV radiation according to increasing doses according to substantially circular spots.

[0015] An experienced practitioner determines for each spot that it reaches the minimum erythemal dose or MED, ie, the UV dose received by the skin that causes erythema to appear.

[0016] Thus, according to the criteria, the MED is the lowest dose of ultraviolet (UV) radiation that produces the first perceptible clear erythema within 16 h to 24 h after exposure to said UV radiation, with regular borders and appearing over the majority of the UV-exposed sites.

[0017] The UV dose corresponding to the MED of the unprotected part (called MEDu) is lower than the UV dose corresponding to the MED of the protected area (called MEDp).

[0018] Therefore, according to this method, the sun protection factor or individual SPF (SPFi) for a given subject corresponds to the ratio between the MED of the protected area and the MED of the unprotected area of ​​that same subject, and SPFi = MEDp / MEDu.

[0019] The SPF of a product is the arithmetic mean of all valid individual SPF values ​​obtained from all subjects.

[0020] The conditions related to the UV exposure source (combined UVA and UVB), exposure zone and subjects are determined by the ISO 24444:2010 standard.

[0021] It should be noted that, according to the standards, the determination of the erythema area used to determine the MED is performed "blindly" by one or more practitioners who are not involved in the application of the product and the exposure of the area under consideration and therefore do not know the dose received. Therefore, this task actually corresponds to the collection of data, subsequently anonymized, aimed at characterizing the product rather than making a diagnosis.

[0022] This testing method is time-consuming and based on a series of visual interpretations, which are inevitably subjective and dependent on the viewing conditions that lead to these interpretations. In addition, the diversity of skin tones, lighting conditions during analysis, and the visual acuity of the practitioner can also affect the results.

[0023] Because this method requires identification of the spot that received the UV dose (actually the minimum dose, corresponding to the appearance of erythema), it is less reproducible. The exposure doses are discrete and incremental, and there may be ambiguity between spots that received an immediate lower dose and spots that received an immediate higher dose.

[0024] Other irritation tests on exposed areas of skin (particularly safety tests involving a series of patches) are based on interpretation and comparison between exposed areas and therefore have an interpretative and subjective component that may affect the results.

[0025] Document FR 3005255 describes a photogrammetric method in which the image of the exposed area is normalized and segmented with reference to a test pattern. The detection of the erythema area is carried out via a learning method. This method limits the influence of environmental conditions (such as lighting, skin color), but still associates an interpretation within the learning range.

[0026] Document US 2014 / 0088380 describes a multispectral imaging system for detecting bruises and pressure ulcers and uses for this purpose an acquisition approach based on a reduced number of spectral bands.

[0027] Document WO 03 / 30707 describes a method for determining the minimal erythemal UV dose by means of multispectral imaging using a spectral band whose intensity of reflected light can be correlated to the amount of melanin in the skin.

[0028] In contrast to the method subject of the present invention, these prior art devices and methods do not use a spectral demixing operation that can extract erythema features in multispectral images. Summary of the invention

[0029] The object of the invention is a device or a system implementing an automatic or semi-automatic method for characterizing and comparing erythema areas, particularly suitable for determining the sun protection factor (SPF) of photoprotective products, limiting the influence of the interpretation part, or even eliminating any interpretation.

[0030] Therefore, the present invention relates to a system for comparing erythema areas, the system comprising: a multispectral camera for acquiring multispectral images of comparison areas; a computer processing device including storage, display, input and calculation means, which is capable of retrieving images acquired by the multispectral camera, including: an image processing program, which is capable of performing spectral unmixing of the acquired images and isolating the erythema component of the image; and a program for comparing these images based on the intensity of the erythema component of each image.

[0031] Thus, the system of the present invention advantageously uses multispectral images to extract components specific to erythema conditions, facilitating the interpretation of the images and making the interpretation of the images reliable.

[0032] The invention is advantageously implemented according to the embodiments and variants disclosed hereinafter, which should be considered individually or according to any technically effective combination.

[0033] Advantageously, the images acquired by the multispectral camera and transmitted to the computer system include metadata related to the conditions under which the erythema zone was created and the conditions under which the images were acquired. Thus, the use of these metadata automates all or part of the process.

[0034] According to a particular embodiment, the system subject of the invention is suitable for determining the individual sun protection factor (SPFi) of a photoprotective product, comparing images (including images corresponding to a plurality of sites exposed to increasing UV doses) and metadata, wherein the metadata includes all or part of the information concerning the test product, the corresponding UV dose at each exposure site and the size of the exposure zone. Therefore, the system subject of the invention is compatible with the scheme disclosed in the ISO 24444:2010 standard.

[0035] Advantageously, according to this particular embodiment, the comparison procedure comprises determining, for each comparison image, a response curve giving the intensity of the erythema component as a function of the UV dose. Determining these response curves limits the variability of the results and, in some alternative embodiments, eliminates the step of visually determining the MED.

[0036] In a variant of the system according to the invention suitable for determining the SPFi of a light protection product, the storage means of the computer device of the system according to the invention comprises a threshold value or a band of erythema intensity depending on the UV dose, and the comparison program determines the SPFi by the intersection of this band or this threshold value with the response curve of each comparison area. This variant allows the SPFi to be determined automatically.

[0037] In another variation, the comparison procedure is performed by determining the similarity coefficient that minimizes the squared deviation between the response curves of the comparison areas. k To determine the SPFi value, one of these curves is assigned the coefficient kThis embodiment may determine the SPFi value based on all sites exposed to the UV dose.

[0038] Therefore, the invention also relates to a method for determining the SPFi of a sunscreen product from two images, each image comprising a plurality of spots obtained by exposing the skin to increasing UV doses, one of the series being obtained on unprotected skin and the other being obtained on skin protected by the product, the two series causing erythema characteristics to appear on at least one spot, the method comprising the following steps:

[0039] a. Collect multispectral images for each series;

[0040] b. spectrally unmixing each image to obtain the intensity of the spectral component corresponding to the erythema for each spot in each series;

[0041] c) from the results of step b) to determine the response curve, giving each series depends on the intensity of the erythema component of these UV doses;

[0042] d. Compare the response curves thus obtained in order to determine the SPFi value.

[0043] According to one embodiment of the method of the invention, step d) comprises displaying the erythema component of the images of the protected area and the unprotected area in false colors, indicating on these images the area corresponding to the minimum erythema dose on each image by means of an input device of a computer device, reporting the area corresponding to the minimum erythema dose ( u n1 ) in order to obtain the appropriate dose u p2 , report the response curve corresponding to the minimum erythema dose ( u p1 ) of the erythema intensity in order to obtain the dose u n2 , through volume ( u p1 + u p2 ) / ( u n1 + u n2 )Sure SPFi value 。 This implementation is fully compliant with the ISO 24444:2010 standard, but the determination of the MED value is more reliable.

[0044] According to another embodiment of the method of the present invention, step d) comprises: f n , fp approximate response curve, and the SPFi is the function f n , f p The mean square deviation between the lowest UV exposure dose and the highest UV exposure dose in the compared area is minimized. k This embodiment can automatically determine the SPFi value by considering all exposure sites.

[0045] Advantageously, step a) comprises acquiring a healthy skin area (without erythema), and step b) uses the acquired healthy skin area to perform spectral unmixing to determine the erythema component.

[0046] The systems and methods subject of the present invention further allow for the automation of the characterization of erythema areas, for which purpose they are advantageously used for screening, taking into account the selection of ingredients constituting products intended for topical application. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The following is a preferred embodiment of the present invention and is not limited to Figures 1 to 7 The present invention is disclosed in these drawings:

[0048] Related to the prior art Figure 1 Schematically represents two zones comprising multiple sites (spots) exposed to increasing UV doses, Figure 1 A is the unprotected area. Figure 1 B is the area protected by the light protection product;

[0049] Figure 2 Schematic representation of the spectrum obtained from an image of a multispectral camera, Figure 2 A is the comparison between non-erythematous skin and irritated skin, Figure 2 B is the isolated spectrum of the erythema response;

[0050] Figure 3 By separating the healthy skin component and the erythema component, an example of a response curve depending on the received UV dose is given;

[0051] Figure 4 shows examples of response curves for the erythema component of protected and unprotected areas depending on the UV dose exposed;

[0052] Figure 5 Explains from Figure 4 Example of determining the SPFi value of a light protection product using the response curve of the light protection product (MED detection of protected and unprotected areas was implemented);

[0053] Figure 6 Explains from Figure 4Example of determining the SPFi value of a photoprotective product using a response curve (implementing erythema intensity thresholds or spectral bands);

[0054] and Figure 7 An example of an embodiment of the system subject of the invention is schematically represented. DETAILED DESCRIPTION

[0055] Related to the prior art Figure 1 The method used to determine the SPFi value for a given subject is described.

[0056] Figure 1 A, unprotected skin zones (101 to 106) of a subject (whose positions are defined by the standard) are exposed to increasing UV doses. According to one example of implementation, the UV dose increases from zone to zone by a factor of 1.25, so that the sixth exposure zone (106) receives a dose that is 3 times the dose received by the first exposure zone (101).

[0057] The UV dose received by each spot is determined by the exposure time of that spot to the UV source.

[0058] Figure 1 B, Perform the same procedure on the same subject on areas (111 to 116) (this time protected by the test product).

[0059] In order to limit the number of tests, the protection index of the product is assumed to be in the form of, for example, the expected minimum protection index.

[0060] Thus, if the assumed minimum protection factor of the product is, for example, 12, the first zone (111) of the protected portion of the test product is exposed to 12 times the dose to which the unprotected first zone (101) is exposed.

[0061] The next zones (112 to 116) are exposed to increasing doses (eg, by a factor of 1.25 from zone to zone).

[0062] In the case of determining the SPF value of a product, the above-mentioned procedure corresponding to the ISO 24444:2010 standard is a prerequisite for implementing the system of the invention, which implements the method for analyzing the results.

[0063] Figure 1 A (according to the prior art and with reference to this example), the fourth exposed area (104) of the unprotected portion is evaluated as the first to develop erythema in the increasing sequence of exposure represented by the dotted arrows in the figure.

[0064] Therefore, this region corresponds to 4 times (1.25 3 )The largest dose, the UV dose corresponding to this fourth zone is MEDu.

[0065] Figure 1 B (still according to the prior art and with reference to this example), the fifth spot (115) of the protected area is evaluated as the first to develop erythema in the exposure increasing sequence, and the UV dose received by this fifth area (115) of the protected part is the MEDp amount.

[0066] Compared to the unprotected portion of the fourth zone (104) and according to this example, this fifth zone (115) receives a UV dose that is 15 times (1.25 x 12) the UV dose to which the latter has been exposed. Therefore, according to this example, MEDp / MEDu = SPFi =15.

[0067] The same operation is performed on at least 10 subjects according to the standard, and the SPF value of the product is the arithmetic mean of the SPFi determined accordingly.

[0068] The system of the present invention performs a multispectral analysis of the exposed area and, according to various implementations, compares the results from this multispectral analysis of the protected and unprotected areas.

[0069] Figure 7 The system of the present invention comprises a multispectral camera (701) or any other equivalent acquisition device capable of acquiring reflection images according to multiple separated wavelengths in a single or multiple shots of a region of interest.

[0070] Thus, unlike conventional digital photographic images that record scene brightness according to three primary colors (RGB), the pixels of an image obtained by a multispectral camera have N primary colors corresponding to the N acquisition wavelengths of the multispectral camera used.

[0071] By adding nuances of color and information that are invisible to the naked eye, multispectral imaging makes it possible to more precisely define reflected light.

[0072] Under the general term of "multispectral camera", a variety of acquisition techniques are possible (with or without implementing scenario-specific lighting).

[0073] Advantageously, the scene being analyzed, i.e. the skin of a subject that has been subjected to UV exposure, is illuminated during the acquisition of the images by the multispectral camera by an adapted light source, this being to ensure calibrated lighting conditions and irradiances for all wavelengths acquired by the camera. By way of example, the light source comprises an integrating sphere with a plurality of light emitting diodes (LEDs) emitting at different wavelengths corresponding to the acquisition wavelengths of the camera used. Thus, the lighting conditions are calibrated and reproducible.

[0074] According to a non-limiting embodiment, the camera collects according to 19 discrete wavelengths ranging from 365 nanometers to 970 nanometers. The shortest wavelength corresponds to the ultraviolet region and the longest wavelength (over 780 nm) corresponds to the infrared region. The intermediate wavelengths correspond to the visible spectrum from ultraviolet to infrared, and the light source includes a plurality of LEDs whose emission spectra are within the collected spectrum.

[0075] As a minimum, the camera's acquisition spectrum and the irradiance spectrum must be acquired based on multiple primary colors distributed between 400 nm and 700 nm (where the reflectance / absorbance peak of hemoglobin lies) and present (with different intensities) on healthy skin (without erythema) and on skin with erythema.

[0076] Thus, the inventors were able to determine that suitable and reproducible results as disclosed in the Examples at the end of this document were obtained with a multispectral acquisition comprising at least 8 wavelengths covering 400 nm to 700 nm.

[0077] Each pixel of the image recorded by the multispectral camera that is subject of the system of the invention is therefore a mixture of more than 8 primary colors, preferably more than 16 primary colors and, according to one example of embodiment, comprises up to 19 primary colors.

[0078] Images acquired directly on the subject's skin and corresponding to each spot exposed to a UV dose produce brightness or reflectance intensity, or alternatively absorbance, where absorbance = 100 - reflectance, and through spectrally unmixed image analysis (described below) the intensity component attributable to the erythema condition can be determined, and therefore characterize the erythema.

[0079] Figure 2 A compares the spectrum (210) of a skin area with fully developed erythema (giving absorbance (202) depending on wavelength (201)) with the spectrum (211) of a so-called healthy area of ​​skin (without erythema).

[0080] The spectrum of the erythema area image is actually a combination of the spectrum of healthy skin and the changes in this spectrum caused by erythema.

[0081] Figure 2 B, The spectrum of this change caused by erythema is obtained by subtracting the spectrum of the healthy skin area from the spectrum of the erythema area (212). This shows a more pronounced absorbance especially for wavelengths close to 580 nm (characteristic of the absorbance of hemoglobin).

[0082] This spectrum (212) is somewhat characteristic of erythema.

[0083] Therefore, the images of the erythema region were analyzed by a spectral unmixing method in order to extract the contribution of the erythema in terms of the abundance of spectral bands according to this feature.

[0084] The spectral unmixing technique is known from the prior art and consists of decomposing the measured spectrum into a set of pure components (also called end Yuan ) and estimate the corresponding proportion (also called abundance) of each of these components. In other words, the spectrum observed in each pixel is assumed to be equal to the combination of the spectral characteristics of a certain number of pure components, healthy skin components, and erythema components, with the goal of evaluating the proportion of each of them.

[0085] According to an exemplary embodiment, the method implemented by the system of the invention uses a linear spectral unmixing technique, making the assumption that the spectrum measured in each pixel corresponds to a weighted average (depending on the abundance) of the spectral characteristics of the pure components (healthy skin and erythema).

[0086] Without going into further detail, spectral unmixing therefore consists of: y = Ma + n The type of equation expressing the measured spectrum depends on the pure components y ,in M is the spectral feature matrix of the pure components, a is the abundance vector, n is a Gaussian perturbation, and the unknown is the abundance vector a .

[0087] In the present case, there are two pure components or end members: healthy skin and erythema. Figure 2 A gives an example of a spectrum corresponding to healthy skin (211), Figure 2 B gives an example of a spectrum corresponding to the erythema (212).

[0088] The solution of this equation is to minimize the residual r， like r = y - Ma The constrained regression algorithm is used to obtain

[0089] According to an implementation example, the method implemented by the system of the present invention uses for this purpose so-called parsimonious calculation methods, such as SUnSAL ( S parse Un mixing by variable S plitting and A ugmented L agrangian [mixing of sparse solutions by variable splitting and augmented Lagrangian] method.

[0090] Applying this method to a multispectral image of skin areas exposed or not exposed to UV and with or without erythema can to some extent classify the pixels of this image into the class of "healthy skin" and the class of "erythema".

[0091] Figure 3 , the above treatment is applied to 6 zones (such as Figure 1 A or Figure 1 as shown in B).

[0092] Image processing can determine for each of the above regions the intensity (302) of the absorbance corresponding to healthy skin (311) and erythema (310). This last curve (310) shows that the relative abundance of the erythema spectral feature increases as the UV dose to which the region is exposed increases.

[0093] The results show the relevance of the applied spectral unmixing, with the intensity corresponding to healthy skin (311) being essentially constant (regardless of the intensity of the erythema).

[0094] The curve (310) corresponding to the absorbance intensity associated with the erythema condition increases with the sequential number of the exposure spots, ie with the UV dose. Furthermore, this curve (310) represents the erythema response of the skin depending on the UV dose received.

[0095] To implement the system of the present invention, only the erythema response curve (310) is required, and the curve (311) corresponding to healthy skin (without erythema) is only Figure 3 denoted in order to illustrate the relevance of the spectral unmixing operation performed, or only used according to an improved implementation.

[0096] Figure 4 , by applying pressure to unprotected skin (e.g. Figure 1 A) of multispectral image acquisition and protected areas (such as Figure 1 B) The skin is collected using a spectral unmixing method, and two intensity curves of absorbance (402) are obtained depending on the UV dose (401). Curve (420) corresponds to the erythema response curve of the unprotected area, which naturally corresponds to the erythema response curve (421) when the UV dose is lower than that of the protected area, while the appearance of erythema in the protected area corresponds to a higher UV dose.

[0097] The deviation between these response curves (420, 421) represents the SPFi of the product being tested.

[0098] The two response curves (420, 421) corresponding to the unprotected area and the protected area, respectively, do not necessarily have exactly the same shape, but have a certain degree of similarity.

[0099] Curve φ n(420) is the response curve of the unprotected area, and curve φ p (421) is the response curve of the protected area.

[0100] These curves represent the absorbance intensity of the erythema component of the skin, which depends on the UV dose to which it has been exposed. Therefore, the curve φ n ( u n ) gives the dependence of UV dose (each spot in the unprotected area i = 1…6 Accepted u n (i) ) of erythema intensity ， φ p ( u p ) gives the UV dose (area protected by the light protection product) for each spot i = 1…6 Accepted u p (i) ) of erythema intensity.

[0101] Assuming that the light protection product has an expected minimum protection index SPF0, then u p (i) = u n (i).SPF 0 。

[0102] To determine SPFi, we assume that the photoprotective product absorbs, reflects or scatters a portion of the UV radiation and only allows 1 / k of this radiation reaches the skin.

[0103] The response of the skin is controlled only by the effective UV dose. This assumption is fully in line with ISO 24444:2010. Therefore, the photoprotective product under test protects and is exposed to a dose u The skin area receives an effective radiation dose u / k .

[0104] In order to implement the system of the present invention in accordance with ISO standards, the UV doses to which the protected and unprotected parts are exposed must be preselected so that at least one erythema corresponding to the definition of the ISO 24444:2010 standard appears in each area within 16 hours to 24 hours after this exposure, that is, a perceptible, clear erythema with regular borders appears on most of the UV exposure sites (spots).

[0105] The system of the present invention determines the coefficients by comparing the response curves (420, 421) according to one or more methods (as disclosed below, alone or in combination).k (which can be interpreted as the similarity coefficient between the two curves) 。

[0106] Figure 5 , a variety of methods can be used for this purpose.

[0107] According to a first example of implementation, the method implemented by the system of the invention remains fully compliant with the ISO 24444:2010 standard and involves visual determination of MEDu and MEDp. Thus, the intensity of the erythema component of the spot corresponding to MEDu is determined on the unprotected area (510) in the same way as the intensity of the MEDp corresponding to the protected area (511). Each of these intensities corresponds to a UV dose, u n1 ，u p1 .

[0108] By reporting the response curve (420) of the unprotected area, corresponding to the intensity (511) of the MEDp identified on the protected area, and by reporting the response curve (421) of the protected part, corresponding to the intensity (510) of the MEDu, the values ​​corresponding to the UV doses can be defined. u n2 and u p2 , which causes a erythema on the unprotected area whose intensity is equivalent to the intensity of the spot corresponding to the protected area where the MEDp is detected, and vice versa.

[0109] In the example of this embodiment, SPFi is, for example, determined by the ratio SPFi = (u p1 + u p2 ) / (u n1 + u n2 ) Given.

[0110] This method, performed in a manner similar to double weighting, improves the accuracy and reproducibility of SPFi determination compared to prior art methods.

[0111] Figure 6 According to another example of an embodiment, a threshold (620) or a band (630) corresponding to the intensity of the presence of significant erythema is determined from statistical data or by learning. The intersection of the threshold (620) or the band (630) with each response curve can be determined. k value and obtain SPFi.

[0112] According to one embodiment, the learning algorithm is pre-optimized and parameterized from a library of reference images annotated with clinical scores of MED.

[0113] Therefore, by referring to Figure 6 , when estimated by threshold (620), SPFI = (u p / u n ) , or in the case of estimation by spectral band (630) SPFi = (u p1 + u p2 ) / (u n1 + u n2 ) .

[0114] Figure 6 Represents a threshold (620) and a spectral band (630), whose characteristics are constant depending on the UV dose (401).

[0115] According to a preferred embodiment, the method implemented by the system subject of the present invention is obtained from the response curve φ of the unprotected part. n ( u n )(420) and the response curve φ of the protected part p ( u p ) = φ p ( ku n )(421)Group determination k The value that minimizes the squared deviation between the two curves k value, which is equivalent to minimizing the following quantity:

[0116]

[0117] in u is the UV dose.

[0118] To do this, the response curve φ n ,φ p Through sigmoid f n 、f p Approximately, the mathematical formula becomes:

[0119]

[0120] Minimize the amount k The value is equal to SPFi.

[0121] Therefore, according to this embodiment, the method implemented by the system subject of the present invention does not require the determination of MEDu and MEDp, or any learning, thus eliminating the uncertainty of this determination, and evaluates the SPFi of the exposed spot group (for both the unprotected part and the protected part).

[0122] Figure 7 Regardless of the implementation, the system of the present invention includes a multispectral camera (701) and a processing device (702) such as a microcomputer.

[0123] The microcomputer conventionally comprises storage means, input and pointing means such as a keyboard and a mouse, display means such as one or more screens and a microprocessor.

[0124] The microcomputer (702) is connected to the multispectral camera (701). According to alternative embodiments, the connection is a wired link (the images acquired by the camera are directly transmitted to the microcomputer, which, according to one example of embodiment, also controls the acquisitions performed by the camera), a wireless connection such as WIFI® or Bluetooth® (the microcomputer and the multispectral camera are members of the same personal network or WPAN), or also via a network connection of the WAN type or via the Internet. This last embodiment allows the acquisition and processing of information to be carried out at a remote site and / or to be temporarily transferred, for example, in order to carry out the acquisition and processing in batches.

[0125] Advantageously, the image transmitted by the multispectral camera includes a metadata file including all or part of the information on: the conditions under which the image was acquired, the product tested in coded form, the subject in coded form (in particular their individual typological angles characterizing the color of their skin), the UV dose corresponding to each exposure zone, the dimensions of the exposure zones. The data related to the subject and the product are entered, for example, during the acquisition of the image, for example by scanning a smart code such as a barcode or a QR code.

[0126] In particular, this information can be used to correlate acquisitions related to unprotected skin and acquisitions related to protected skin within a microcomputer via analysis of these data, and then grouping the processing results in order to calculate the SPF.

[0127] Information related to the subject and product was coded to allow a “blind” analysis of the results according to the criteria.

[0128] The microcomputer comprises, in its storage means or via a network access shared application, a program for processing the images coming from the multispectral camera in order to carry out a spectral unmixing of the images (preferably via the SUnSAL method), isolating the component due to erythema in said images.

[0129] The classification process can determine the response curve corresponding to each acquisition from the results of the image processing process (420, 421, Figures 4 to 6 ).

[0130] After this processing, the aforementioned microprogram or an additional display program can display the image (710) in false color (showing the intensity of the erythema component on each spot of the acquired area) and compare the results obtained on the protected area (711) and the unprotected area (712).

[0131] In addition, display of this image can facilitate determination of MEDu and MEDp by visually enhancing the erythema component. Image processing not only returns intensity information of the erythema by displaying false colors, but also returns spatial information within the range of the erythema area within each spot, which helps determine the area corresponding to MEDu and MEDp.

[0132] Therefore, according to the first embodiment, the analyst first points out the region (spot) corresponding to MEDu, and then points out the region (spot) corresponding to MEDp.

[0133] Also according to this first embodiment, after verification, the program extracts the data files corresponding to the UV doses associated with these areas and calculates the UV doses by using the data files as described above. Figure 5 The presented response curve is used to calculate the SPFi and then the obtained values ​​are stored in a result file.

[0134] According to another embodiment, from a threshold value or from a spectral band (such as reference Figure 6 SPFi is determined by the computer (see Figure 700 ) for determining the SPFi. For this purpose, a threshold value or a spectral band, or more advantageously, a plurality of threshold values ​​or a plurality of spectral bands are stored, for example, in a storage device of a computer device ( 701 ) associated with a skin color feature. The selection of the threshold value or the spectral band is performed based on metadata transmitted with the image.

[0135] According to another embodiment, the response curves of the protected and unprotected parts are approximated by sigmoids, and the SPFi of the product is calculated by minimizing the difference between the response curve of the protected area and the response curve of the unprotected area (assigned k coefficient, k The embodiment can automatically determine SPFi by image analysis.

[0136] By advantageously combining some or all of these embodiments and providing multiple SPFi values ​​for a given acquisition, according to the combined example of this embodiment, if the multiple values ​​corresponding to the various methods are contained within an interval, such as within + / - 10% centered on the average value of the multiple values, the obtained SPFi value is considered valid.

[0137] All valid results corresponding to the tested product are processed and stored in the result file and are used to determine the SPF of the product.

[0138] If, in addition to the area exposed to irritation, the method implemented by the system of the invention is improved, the acquisition of a healthy skin area close to the test area is performed, regardless of the cause of this irritation (UV exposure or product during topical application). This acquisition is therefore advantageously used to determine the spectral component of erythema, as referenced Figure 2 disclosed, or for controlling the effectiveness of the response curve, as referenced Figure 3 Disclosed.

[0139] Advantageously, the acquisitions made on healthy skin are associated with the test pattern.This embodiment makes it possible to determine the color of the subject's skin, in particular by processing the images.

[0140] The examples of the embodiments disclosed above are mainly disclosed within the scope of determining the sun protection factor of a photoprotective product, however, all or part of the disclosed principles and embodiments can be used to characterize the erythema area within the scope of other skin irritation tests (comparison of the erythema area or comparison between the irritated area and the healthy area).

[0141] The possibility of automatically characterizing the area subjected to stimulus testing by means of the device and method of the invention enables certain applications to process a large number of tests quickly and reliably.

[0142] Thus, the method and system of the present invention are advantageously used in systematic testing within the context of screening ingredients that may go into the composition of topical application products, these tests being related to their irritation potential or their protective effect.

[0143] Examples

[0144] To test the validity of the method, a study was conducted on 10 subjects for determining the SPFi value of a product (reference P2 in ISO 24444:2010 standard).

[0145] Therefore, each subject was exposed to UV in 2 areas of the back, the first area was not applied with a photoprotective agent, and the second area was applied with a photoprotective product P2. The product consists of 3 phases, Phase 1: lanolin 4.5%; cocoa seed butter 2.0%; glyceryl monostearate 3.0%; stearic acid 2.0%; 4-(dimethylamino)benzoic acid 2-ethylhexyl ester PABA (CAS 21245-02-3) 7.0%; benzophenone-3 (CAS 131-57-7) 3.0%; Phase 2: water 71.6%; sorbitol (70% of liquid) 5.0%; triethanolamine 1.0%; methylparaben 0.3%; propylparaben 0.1%; Phase 3: benzyl alcohol 0.5%. The above data are all mass fractions (%).

[0146] Each exposure zone consisted of 6 spots with a dose escalation factor of 1.25 (according to the standard). For each subject, zone 2 (with photoprotective product applied) had been exposed to a dose 14 times greater than zone 1. The exposure system used was a UV SOLARLIGHT MULTIPORT BERGER (model 601 at 300 W) whose spectrum complies with the standard.

[0147] 24 hours after exposure, multispectral acquisition was performed on each area of ​​each subject at 19 wavelengths between 365 nm and 970 nm with the aid of the device subject of the present invention (including a multispectral camera), and the SPFi value corresponding to each subject was determined by comparing the response curves of the protected part and the unprotected part after spectral unmixing and erythema component extraction.

[0148] The wavelength distribution of camera collection is shown in Table 1.

[0149] Band number Band name Wavelength [nm] 1 UVA 365 2 Purple 405 3 Indigo 435 4 blue 450 5 blue 470 6 Teal 505 7 green 525 8 yellow 570 9 amber 590 10 red 630 11 red 645 12 red 660 13 Dark Red 700 14 Near Infrared 780 15 Near Infrared 850 16 Near Infrared 870 17 Near Infrared 890 18 Near Infrared 950 19 Near Infrared 970

[0150] The FPSi calculation results for each subject are shown in Table 2 below:

[0151] Subject number FPSi 1 11.8 2 17.4 3 7.2 4 18.2 5 23.4 6 16.4 7 15.4 8 17.6 9 15.4 10 17.8

[0152] The arithmetic mean of SPFi represents the SPF value of the product, that is, the SPF is 16.06.

[0153] This value of 16.06 corresponds to the standard reference value (16.1) for product P2; the standard indicates in its Annex C that the average SPF value is 16.1 and is between 13.7 and 18.5.

[0154] This example illustrates the relevance of the method subject of the invention for automatically determining the SPFi value.

Claims

1. A method for determining the SPFi value of a photoprotective product from two series of images, each series comprising a plurality of spots obtained by exposing the skin to increasing UV doses, one series of spots being obtained on unprotected skin and the other series of spots being obtained on skin protected by the product, said two series causing erythema characteristics to appear on at least one of said spots, the method comprising the following steps: a. collecting each of the series of multispectral images (711, 712) according to 19 discrete wavelengths ranging from 365 nanometers to 970 nanometers; b. performing spectral unmixing on each of the multispectral images to separate the spectral component corresponding to erythema in the image, so as to obtain the intensity (402) of the spectral component (212) corresponding to the erythema for each of the spots in each series; c. determining from the result of step b) a response curve (420) of an unprotected area and a response curve (421) of a protected area, giving the intensity of the spectral component corresponding to erythema for each of said series depending on the UV dose; d. Comparing the response curve (420) of the unprotected area and the response curve (421) of the protected area thus obtained to determine the SPFi value; wherein step d) comprises calculating the SPFi value by sigmoid f n and sigmoid f p The response curve (420) of the unprotected area and the response curve (421) of the protected area are respectively approximated, and the SPFi value is the function sigmoid f n and sigmoid f p The mean square deviation between the lowest UV exposure dose and the highest UV exposure dose in the compared areas is minimized. k value.

2. The method according to claim 1, wherein: Step a) comprises acquiring a healthy skin area without erythema, and step b) uses this acquisition to perform spectral unmixing to determine the spectral components corresponding to said erythema (212).

3. The method according to claim 1, wherein: The multispectral image comprises at least 8 primary colors per pixel corresponding to wavelengths covering 400 nm to 700 nm.

4. A system for implementing the method according to claim 1, comprising a multispectral camera (701) for acquiring multispectral images (711, 712) of skin exposed areas according to 19 discrete wavelengths ranging from 365 nanometers to 970 nanometers, a computer processing device (702) including storage, display, input and calculation means capable of retrieving images acquired by the multispectral camera, characterized in that The computer processing device runs: an image processing program capable of spectrally unmixing the acquired image and isolating the spectrum of the spectral component (212) of the image corresponding to erythema; and a comparison procedure for comparing the images based on the intensity of the spectral component of each image corresponding to erythema.

5. The system according to claim 4, wherein: The images acquired by the multispectral camera (701) and transmitted to the computer processing device (702) include metadata related to the conditions under which the erythema area was created and the conditions under which the images were acquired.

6. The system according to claim 4, wherein: The storage device of the computer processing device (702) stores a threshold value (620) or a spectrum band (630) of erythema intensity depending on the UV dose (401), and wherein the comparison program determines the SPFi value by the intersection of the spectrum band (630) or the threshold value (620) with the response curve (420) of the unprotected area and the response curve (421) of the protected area of ​​each comparison area.

7. Use of a system according to claim 4 for implementing a method according to claim 1, for screening to select ingredients constituting a product for topical application.

Images