DEVICE FOR CHARACTERIZING AND COMPARISON OF ERYTHEMA ZONES

Description

[0001] The invention relates to a device for characterizing and comparing erythemal areas of the skin. The invention is more particularly, but not exclusively, suitable for determining a sun protection index, or Sun Protection Factor (SPF), of a photoprotective product intended to be applied to the skin. Prior art

[0002] The development and evaluation of a cosmetic product requires numerous tests, including skin irritation tests aimed, for example, at verifying the safety of a product or its protection index.

[0003] These tests are conducted by exposing human or animal skin to the product or to a stress such as exposure to ultraviolet radiation, using different methods, then detecting and characterizing erythemal areas on the regions thus exposed.

[0004] The characterization of erythemal areas is based on a comparison, either between the said erythemal area and healthy skin (without erythema) or between two erythemal areas subjected to different conditions.

[0005] The detection of erythema and its characterization is based on the interpretation of erythemal conditions by a practitioner, for example a dermatologist.

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

[0007] A photoprotective product is a product: cream, milk, balm or powder containing any component that can absorb, reflect or diffuse UV rays, which is intended to be placed in contact with human skin. Such a product acts by reflecting, absorbing or diffusing radiation external to the skin and has no therapeutic action on it.

[0008] Alternatively, the photoprotective product is a textile.

[0009] The sun protection factor determines the increase in UV exposure time provided by the application of the product before this exposure produces erythema, more commonly referred to as "sunburn".

[0010] Thus, if a subject exposed to a given UV radiation without protection develops erythema after 5 minutes of exposure to this radiation, the same subject, protected by the application of a photo-protective product with an SPF of 10, can be exposed to this same radiation for 50 minutes before developing erythema.

[0011] Determining the sun protection factor (SPF) of such a product involves in vivo testing involving exposure of a subject's protected and unprotected skin to ultraviolet radiation under controlled conditions.

[0012] Thus, the determination of the sun protection factor of a product applied to the skin is defined by the ISO 24444:2010 standard, which consists of an in vivo test carried out on the skin of several subjects.

[0013] According to this test, for a given subject, a first part of the skin, unprotected, that is to say without application of the photo-protective product, is exposed, according to approximately circular spots, to UV radiation according to increasing doses.

[0014] A second part, adjacent to the first part, is covered by the protective product and exposed, in approximately circular spots, to UV radiation also in increasing doses.

[0015] An experienced practitioner determines for each of these parts the spot for which the Minimum Erythemal Dose, or MED, is reached, that is to say the spot for which the UV dose received by the skin causes erythema to appear.

[0016] Thus, according to the standard, DEM is the smallest dose of ultraviolet (UV) radiation that produces the first unambiguous perceptible erythema, with regular edges and appearing over most of the UV exposure site, within 16 to 24 hours of exposure to said UV radiation.

[0017] The UV dose corresponding to the DEM of the unprotected part, called DEMn, is lower than the UV dose corresponding to the DEM of the protected area, called DEMp.

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

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

[0020] The conditions relating to the source of UV exposure, mixing UVA and UVB, to the exposed areas and to the selection of subjects are determined by the ISO 24444:2010 standard.

[0021] It should be noted that, according to the standard, the determination of the erythemal area used for the determination of the DEM is carried out "blindly" by one or more practitioners who have not participated in the application of the product and the exposure of the area considered, therefore not knowing the dose received. This task therefore corresponds to a collection of data, subsequently processed anonymously, aimed at characterizing a product and not at making a diagnosis.

[0022] This test method is time-consuming to implement and relies on a series of visual interpretations, which are necessarily subjective and dependent on the observation conditions leading to these interpretations. The variety of skin colors, the lighting conditions during the analysis and the visual acuity of the practitioner are all factors, among others, likely to influence the results.

[0023] Reproducibility is all the more delicate since the method requires determining the spot that received a UV dose, in practice the smallest dose, corresponding to the appearance of erythema. Since the exposure doses are discrete and incremental, there may be ambiguity with the spot that received an immediately lower dose and the spot that received an immediately higher dose.

[0024] Other skin irritation tests involving different exposed areas, in particular safety tests involving a series of patches, rely on interpretations and comparisons between exposed areas and therefore also have an interpretative and subjective element, which may introduce variability into the results.

[0025] Document FR3005255 describes a photogrammetric method in which an image of the exposed areas is normalized and segmented, with reference to a target. The detection of the erythemal area is carried out using a learning method. This method limits the influence of environmental conditions (lighting, skin color, etc.) but remains linked to an interpretation within the framework of learning.

[0026] Document US2014 / 0088380 describes a multispectral imaging system used for the detection of bruises and pressure ulcers, and for this purpose uses acquisition according to a reduced number of spectral bands.

[0027] Document WO 03 / 30707 describes a method for determining a minimum erythemal UV dose, by multispectral imaging, using a spectral band whose intensity of reflected light is correlatable to the quantity of melanin in the skin.

[0028] Unlike the method which is the subject of the invention, these devices and methods of the prior art do not use a spectral unmixing operation making it possible to extract an erythemal signature in the multispectral image. Statement of the invention

[0029] The subject of the invention is a device or system, implementing an automated or semi-automated method, for the characterization and comparison of erythemal areas, in particular adapted to the determination of the sun protection factor (SPF) of a photo-protective product making it possible to limit the influence of the interpretative part, or even to free oneself from any interpretation.

[0030] Thus, the invention relates to a system for comparing erythemal areas, meeting the definition of claim 5 and capable of meeting the characteristics of claims 6 to 8, comprising a multispectral camera for acquiring a multispectral image at discrete wavelengths between 365 nm and 970 nm of the areas compared, a computer processing device comprising memory, display, input and calculation means, capable of recovering the images acquired by the multispectral camera, comprising an image processing program capable of performing spectral unmixing of the acquired images and isolating the erythemal component of said images, said spectral components corresponding to the modification, attributable to erythema, of the spectrum of the skin compared to the spectrum of healthy skin, and a program for comparing the images on the basis of the intensity of the erythemal components of each image.

[0031] Thus, the system which is the subject of the invention advantageously uses a multispectral image to extract the components specific to an erythemal situation and thus facilitate and make the interpretation of the images more reliable.

[0032] The invention is advantageously implemented according to the embodiments and variants set out below, which are to be considered individually or in any technically effective combination.

[0033] Advantageously, the images acquired by the multispectral camera and transmitted to the computer system include metadata relating to the conditions for creating the erythemal zones and the conditions for acquiring the images. Thus, the use of this metadata makes it possible to automate all or part of the processing.

[0034] According to a particular embodiment, the system that is the subject of the invention is suitable for determining the individual sun protection factor (IFP) of a photoprotective product, the compared images comprising the images of a plurality of sites corresponding to exposures to increasing UV doses and the metadata comprising all or part of the information relating to the product tested, to the UV doses corresponding to the different exposed sites and to the size of the exposed areas. Thus, the system that is the subject of the invention is compatible with the protocol set out in the ISO 24444:2010 standard.

[0035] Advantageously, according to this particular embodiment, the comparison program comprises the determination for each compared image of a response curve giving the intensity of the erythemal component as a function of the UV dose. The determination of these response curves makes it possible to limit the variability of the results and in certain embodiment variants to eliminate the step of visual determination of the DEM.

[0036] In a variant of the system which is the subject of the invention adapted to the determination of the FPSi of a photoprotective product, the memory means of the computer device of the system which is the subject of the invention comprise a threshold or a band of erythemal intensity which is a function of the UV dose, and the comparison program determines the FPSi by the intersections of this band or this threshold with the response curves of each compared zone. This variant allows the automatic determination of the FPSi.

[0037] In another variant, the comparison program determines the FPSi by determining a homothety factor k minimizing the quadratic deviation between the response curves of the compared areas, one of the curves being affected by the factor k. This embodiment makes it possible to determine the FPSi on the basis of all the sites exposed to the UV doses.

[0038] Thus, the invention also relates to a method, meeting the definition of claim 1 and capable of meeting the characteristics of claims 2 to 4, for determining the SPF of a sunscreen product from 2 images each comprising a plurality of spots obtained by exposing skin to increasing UV doses, one of the series being obtained on unprotected skin and the other on skin protected by the product, both series showing an erythemal character on at least one of the spots, comprising the steps consisting of: a. acquiring a multispectral image at discrete wavelengths between 365 nm and 970 nm, of each of the series; b. carrying out spectral unmixing for each of the images to isolate in said images the spectral components corresponding to erythema, said spectral components corresponding to the modification, attributable to erythema, of the skin spectrum compared to the spectrum of healthy skin, so as to obtain for each spot of each series the intensity of the spectral components corresponding to erythema; c. determining from the result of step b) a response curve giving the intensity of the erythemal components as a function of the UV dose for each of the series; d. comparing the response curves thus obtained to determine the FPSi.

[0039] According to an embodiment of the method which is the subject of the invention, step d) comprises the display in false colors of the erythemal components of the images of the protected area and of the unprotected area, and the pointing on these images, by means of the input means of the computer device, of the areas corresponding to the minimum erythemal doses on each of the images, reporting the erythemal intensity corresponding to the minimum erythemal dose ( the n1 ) on the protected area response curve to obtain the dose up 2 corresponding, report the erythemal intensity corresponding to the minimal erythemal dose ( the p1 ) on the response curve of the unprotected area to obtain the dose the n2, determine FPSi by the quantity (the p1 + the p2 ) / (the n1 + the n2). This embodiment is fully compliant with ISO 24444:2010 but the determination of the DEM is more reliable.

[0040] According to another embodiment of the method which is the subject of the invention, step d) comprises the approximation of the response curves by sigmoids fn , fp and the FPSi is the value of k minimizing the mean square deviation between functions fn , fp over a range from the lowest UV dose to the highest UV dose of exposure of the compared areas. This embodiment allows automatic determination of the FPSi by taking into account all exposure sites.

[0041] Advantageously, step a) includes the acquisition of an area of ​​healthy skin (without erythema) and step b) of spectral unmixing uses this acquisition for the determination of the erythemal components.

[0042] The system and method that are the subject of the invention also make it possible to automate the characterization of erythemal areas; for this purpose, they are advantageously used to carry out screening with a view to selecting the ingredients constituting a product intended for topical application. Summary description of the drawings

[0043] The invention is set out below according to its preferred embodiments, in no way limiting and with reference to the figures 1 to 7 in which: there figure 1 relating to the prior art schematically represents two zones comprising a plurality of sites (spots) exposed to increasing UV doses, figure 1A an unprotected area, figure 1B an area protected by a photo-protective product; figure 2 schematically represents the envelope of the spectra obtained from the images of the multispectral camera, figure 2A the comparison between skin without erythema and irritated skin, figure 2B the isolated spectrum of the erythemal response; figure 3 gives an example of response curves as a function of the UV dose received, separating the healthy skin component and the erythemal component; figure 4 shows an example of response curves of the erythemal components of a protected area and an unprotected area as a function of the UV exposure dose; figure 5 illustrates an example of determining the FPSi of a photoprotective product from the response curves of the figure 4 , implementing DEM detection for the protected area and for the unprotected area; figure 6 illustrates an example of determining the FPSi of a photoprotective product from the response curves of the figure 4 , implementing a threshold or band of erythemal intensity; and the figure 7 schematically represents an example of an embodiment of the system which is the subject of the invention. Best way to carry out the invention

[0044] There figure 1 , relating to the prior art illustrates the method of determining an FPSi value for a given subject.

[0045] Figure 1A , areas of the subject's skin (101 to 106), without protection, the location of which is defined by the standard, are exposed to increasing UV doses. According to an exemplary embodiment, the UV dose increases by a factor of 1.25 from one area to the next, so that the 6th exposed area (106) receives a dose 3 times greater than that of the first exposed area (101).

[0046] The UV dose received by each spot is determined by the time the spot is exposed to the UV source.

[0047] Figure 1B , the same operation is carried out, on the same subject, on areas (111 to 116) this time protected by the tested product.

[0048] To limit the number of tests, the product's protection index is assumed, for example in the form of an expected minimum protection index.

[0049] Thus, if the presumed minimum protection factor of the product is, for example, 12, the first zone (111) of the part protected by the tested product is exposed to a dose 12 times higher than that to which the first zone (101) of the unprotected part was exposed.

[0050] The following areas (112 to 116) are exposed to increasing doses, for example by a factor of 1.25 from one area to another.

[0051] The procedure described above and which corresponds to the ISO 24444:2010 standard is a prerequisite for the implementation of the system which is the subject of the invention and which implements a method for analyzing the results, in the case of determining the SPF of a product.

[0052] Figure 1A, according to the prior art and with reference to this example, the fourth exposed area (104) of the unprotected part is evaluated as being the first, in the increasing order of exposure represented by the broken arrow in this figure, to show erythema.

[0053] This zone thus corresponds to a dose 4 times higher (1.25 3< ) than the dose received for the first zone (101), and the UV dose corresponding to this fourth zone is the quantity DEMn.

[0054] Figure 1B , still according to the prior art and with reference to this example, the fifth spot (115) of the protected area is evaluated as being the first in the increasing order of exposure to cause erythema to appear, and the UV dose received by this fifth zone (115) of the protected part is the quantity DEMp.

[0055] Compared to the fourth zone (104) of the unprotected part and according to this example, this fifth zone (115) received a UV dose 15 times higher than that to which the latter was exposed (1.25x12). Also according to this example DEMp / DEMn= FPSi=15.

[0056] The same operation is carried out on at least 10 subjects, in accordance with the standard, and the FPS value of the product is the arithmetic mean of the FPSi thus determined.

[0057] The system which is the subject of the invention implements a multispectral analysis of the exposed areas and compares, according to the methods of the different implementation modes, the results from this multispectral analysis, for the protected area and the unprotected area.

[0058] Figure 7, the system which is the subject of the invention comprises a multispectral camera (701), or any other equivalent acquisition device, capable of acquiring, in a single or several shots of the area of ​​interest, a reflectance image according to several isolated wavelengths.

[0059] Thus, unlike a classic digital photographic image which records the luminances of a scene according to three primaries (RGB), the pixels of the image obtained by the multispectral camera have N primaries corresponding to the N acquisition wavelengths of the multispectral camera used.

[0060] Multispectral imaging allows for more precise definition of reflected light by providing color nuances and information invisible to the naked eye.

[0061] Under the generic term "multispectral camera" several acquisition technologies are possible, implementing specific or non-specific lighting of the scene.

[0062] Advantageously, the analyzed scene, in this case the skin of the subject on which the UV exposure was carried out, is illuminated by a suitable light source during the acquisition of the image by the multispectral camera, in order to ensure calibrated lighting conditions and illumination by all the wavelengths acquired by the camera. For example, the light source comprises an integrating sphere with a plurality of light-emitting diodes (LEDs), emitting in different wavelengths, corresponding to the acquisition wavelengths of the camera used. Thus, the lighting conditions are calibrated and always reproducible.

[0063] According to a non-limiting exemplary embodiment, the camera performs an acquisition according to 19 discrete wavelengths, ranging from 365 nanometers to 970 nanometers. The shortest wavelength corresponds to the ultraviolet range and the longest wavelengths (above 780 nm) to the infrared range. The intermediate wavelengths correspond to the visible light spectrum, from violet to red, and the lighting source comprises a plurality of LEDs whose emission spectrum is within the acquisition spectrum.

[0064] At a minimum, the camera acquisition spectrum and the illumination spectrum must allow acquisition according to several primaries distributed between 400 nm and 700 nm, the zone in which the hemoglobin reflectance / absorbance peaks are located, and which appear (at different intensities) both on healthy skin (without erythema) and on skin with erythema.

[0065] The inventors were thus able to determine that appropriate and reproducible results as set out in the example at the end of this document were obtained with a multispectral acquisition comprising at least 8 wavelengths covering 400nm to 700nm.

[0066] Thus, each pixel of an image recorded by the multispectral camera of the system which is the subject of the invention is a mixture of more than 8 primaries, preferably more than 16 primaries, and according to an exemplary embodiment, comprises up to 19 primaries.

[0067] The image of each spot, acquired directly on the subject's skin, and corresponding to exposure to a UV dose, is translated into a luminance or reflectance intensity, or alternatively an absorbance, with absorbance=100-reflectance, and the analysis of the image by spectral unmixing, as described below, makes it possible to determine the part of this intensity attributable to an erythemal situation and consequently, to characterize the erythema.

[0068] There figure 2A compares the envelope of the spectrum (210) of an area of ​​skin with well-developed erythema, giving the absorbance (202) as a function of the wavelength (201), with the envelope of the spectrum (211) of an area of ​​so-called healthy skin (without any erythema).

[0069] The spectrum of the erythemal area image is actually the combination of the spectrum of healthy skin, and a modification of this spectrum by the presence of erythema.

[0070] Figure 2B , by subtracting from the spectrum of the erythemal zone the spectrum of the healthy skin zone, the spectrum (212) of this modification attributable to erythema is obtained. This notably presents a more marked absorbance for wavelengths close to 580 nm, characteristic of the absorbance of hemoglobin.

[0071] This spectrum (212) constitutes in some way the signature of an erythema.

[0072] Thus, the image of an erythemal area is analyzed by spectral unmixing processing in order to extract the contribution of the erythema in terms of abundances in the spectral bands in accordance with this signature.

[0073] Spectral unmixing techniques are known from the prior art and consist of decomposing the measured spectrum into a set of pure components (also called endmembers ) and to estimate the respective proportions (also called abundances) of each of these components. In other words, the spectrum observed at each pixel is assumed to be equal to a combination of the spectral signatures of a number of pure components, components of healthy skin and components of erythema, the aim being to assess the proportion of each of its components.

[0074] According to an exemplary embodiment, the method implemented by the system which is the subject of the invention uses a linear spectral unmixing technique, assuming that in each pixel the measured spectrum corresponds to the weighted average (as a function of the abundances) of the spectral signatures of the pure components (healthy skin and erythema).

[0075] So, without going into more details, spectral unmixing consists of expressing a measured spectrum y as a function of the pure components by an equation of type y=But + n, Or M is the matrix of spectral signatures of the pure components, a the vector of abundances and n a Gaussian perturbation, the unknown being the vector of abundances a.

[0076] In the case of the invention, there are two pure components or members: healthy skin and erythema. An example of a spectrum corresponding to healthy skin (211) is given figure 2Aand an example of a spectrum corresponding to erythema (212) is given figure 2B .

[0077] The solution to the equation is obtained by a constrained regression algorithm aimed at minimizing the residuals r such as r = y-Ma.

[0078] According to an exemplary implementation, the method implemented by the system which is the subject of the invention uses for this purpose a so-called parsimonious algorithmic method, such as the SUnSAL method, acronym for “Sparse Do not mixing by variable S pleating and A augmented L agrangian ».

[0079] The application of this method to a multispectral image of an area of ​​skin exposed or not to UV and presenting or not erythema allows, in a way, to classify the pixels of this image into a “healthy skin” class and an “erythema” class.

[0080] Figure 3, the treatment described above is applied to 6 areas that have been exposed to UV rays according to increasing doses from area 1 to area 6, as shown in the figure 1A or the figure 1B .

[0081] Image processing makes it possible to determine for each of these areas the intensity of the absorbance (302) corresponding to healthy skin (311) and erythema (310). This last curve (310) shows the increasing relative abundance of the spectral signature of erythema when the UV dose to which the area has been exposed increases.

[0082] This result shows the relevance of the applied spectral unmixing, the intensity (311) corresponding to healthy skin being substantially constant whatever the intensity of the erythema.

[0083] The curve (310) corresponding to the intensity of the absorbance associated with an erythemal situation increases with the order number of the exposed spot, that is to say with the UV dose. Also, this curve (310) is representative of the erythemal response of the skin as a function of the UV dose received.

[0084] For the implementation of the system which is the subject of the invention, only the erythemal response curve (310) is necessary, the curve (311) corresponding to healthy skin (without erythema) not being represented on the figure 3 only to show the relevance of the spectral unmixing operation carried out, or is only used according to perfected implementation modes.

[0085] Figure 4 , by applying the spectral unmixing method on a multispectral image acquisition carried out on unprotected skin, as in Fig. 1A, and on an acquisition carried out on a protected area as in Fig. figure 1B, two curves of absorbance intensity (402) are obtained as a function of the UV dose (401), the curve (420) corresponding to the erythemal response curve for the unprotected area, naturally corresponding to UV doses lower than those of the curve (421) of erythemal response of the protected area, the erythema appearing in the protected area for higher UV doses.

[0086] The difference between these response curves (420, 421) is representative of the FPSi of the tested product.

[0087] The two response curves (420, 421) corresponding respectively to the unprotected zone and to the protected zone, do not necessarily have exactly the same shape, but exhibit a certain degree of homothety.

[0088] The curve ϕ n (420) is the response curve of the unprotected area, and the curve ϕ p (421) is the response curve of the protected area.

[0089] These curves represent respectively the intensity of the absorption of the erythemal component of the skin as a function of the UV dose to which it was exposed. Thus the curve ϕ n ( do not ) gives the intensity of the erythema as a function of the UV dose, one (one), received by each spot i=1 ... 6 of the unprotected area, and ϕ p ( up ) gives the intensity of the erythema as a function of the UV dose, up (i), received by each spot i=1 ... 6 of the area protected by the photoprotective product.

[0090] Assuming the photoprotective product has an expected minimum protection factor of SPF 0, up (i)=un (i).FPS 0 .

[0091] To determine the SPF, it is considered that the photoprotective product absorbs, reflects or diffuses part of the UV radiation and leaves only a proportion 1 / k of this radiation reaching the skin.

[0092] The skin's response is solely controlled by the UV dose actually received. This assumption is fully consistent with ISO 24444:2010. Thus, an area of ​​skin protected by the tested photoprotective product and exposed to a dose u receives an effective dose of radiation u / k.

[0093] For the implementation of the system which is the subject of the invention in accordance with the ISO standard, the UV doses to which the protected part and the unprotected part are exposed must have been previously selected so as to cause at least one erythema corresponding to the definition of the ISO 24444:2010 standard to appear in each of these areas, that is to say a perceptible, unambiguous erythema, presenting regular edges and appearing on the major part of the site (spot) of UV exposure, within 16 hours to 24 hours following this exposure.

[0094] The system which is the subject of the invention determines the factor k, interpretable as a homothety factor between the two curves, by comparing the response curves (420, 421) according to one or more methods as set out below, taken alone or in combination.

[0095] Figure 5 , several methods can be used for this purpose.

[0096] According to a first example of implementation, the method implemented by the system which is the subject of the invention remains in all respects compliant with the ISO 24444:2010 standard and involves the visual determination of the DEMn and the DEMp. Thus, the intensity (510) of the erythemal components of the spot corresponding to the DEMn is determined on the unprotected area, as is the intensity (511) corresponding to the DEMp for the protected area. Each of these intensities corresponds to a UV dose, the n1, the p1 .

[0097] By plotting on the response curve of the unprotected area (420) the intensity (511) corresponding to the DEMp identified on the protected area, and by plotting on the response curve of the protected part (421) the intensity (510) corresponding to the DEMn, it is possible to define the values the number 2 And the p2 , corresponding respectively to the UV dose, leading on the unprotected area to an erythema of intensity equivalent to that corresponding to the spot in the protected area for which the DEMp is detected and vice versa.

[0098] In this example implementation FPSi is given, for example, by the ratio FPSi = (the p1 + the p2 ) / (the n1 + the n2).

[0099] Compared to the prior art method, this method, which proceeds in a way by double weighing, improves the precision and reproducibility of the determination of FPSi.

[0100] Figure 6, according to another example of implementation, a threshold (620) or a band (630) of intensity corresponding to the presence of a significant erythema are determined from statistical data or by learning. The intersections of the threshold (620) or of this band (630) with each of the response curves make it possible to determine a value of k and thus obtain FPSi.

[0101] According to one embodiment, the learning algorithm is previously optimized and parameterized from a reference image bank annotated with the clinical scores of the DEMs.

[0102] So, referring to the figure 6 , FPSI= ( up / do not ) in the case of an estimation by a threshold (620) or FPSi = (the p1 + the p2 ) / (u n1 +u n2 ) in the case of an estimation by a band (630).

[0103] There figure 6 represents a threshold (620) and a band (630) whose characteristics are constant as a function of the UV dose (401).

[0104] According to a preferred embodiment, the method implemented by the system which is the subject of the invention determines the value of k from the set of response curves, ϕ n ( do not ) of the unprotected part (420) and ϕ p ( up ) = ϕ p ( with you ) of the protected part (421), or the value of k which minimizes the quadratic difference between the 2 curves, which amounts to minimizing the quantity: ∫ min u max u φ n u − φ p k . , u 2 . du

[0105] Or u is the UV dose.

[0106] To perform this operation, the response curves ϕ n , ϕ p are approximated by sigmoids fn , fp and the mathematical formulation becomes: ∫ min u max u f n u − f p k . u 2 . du

[0107] The value of k minimizing this quantity is equal to FPSi.

[0108] Thus, according to this embodiment, the method implemented by the system which is the subject of the invention does not require the determination of DEMn and DEMp, or any learning, thus overcoming uncertainties in this determination and carrying out the evaluation of the FPSi on all the exposed spots, both of the unprotected part and of the protected part.

[0109] Figure 7 , whatever the embodiment, the system which is the subject of the invention comprises a multispectral camera (701) and processing means (702) for example in the form of a microcomputer.

[0110] Said microcomputer conventionally comprises memory means, input and pointing means such as a keyboard and a mouse, display means such as one or more screens and a microprocessor.

[0111] The microcomputer (702) is linked electronically to the multispectral camera (701). According to alternative embodiments, this connection is a wired connection, the images acquired by the camera being directly transmitted to the microcomputer, which according to an exemplary embodiment also controls the acquisition carried out by the camera, a wireless connection such as WIFI ®< or Bluetooth ®< the microcomputer and the multispectral camera being members of the same personal network or WPAN, or via a WAN type network connection or a connection via the internet network. This latter embodiment makes it possible to carry out the acquisition and processing of information in remote sites and / or in a time-shifted manner so as, for example, to carry out acquisitions and processing in batches.

[0112] Advantageously, the images transmitted by the multispectral camera include a metadata file comprising all or part of the information concerning the conditions of acquisition of the image, the product tested, in coded form, the subject in coded form, in particular his individual typological angle characterizing the color of his skin, the UV doses corresponding to the different exposed areas, the dimension of the exposed areas. The data relating to the subject and the product are for example entered during the acquisition of the images for example by scanning a smart code, such as a bar code or a QR code.

[0113] This information allows, in particular, via the analysis of this data, to associate within the microcomputer an acquisition relating to unprotected skin and an acquisition relating to protected skin, then to group the treatment results in order to calculate the SPF.

[0114] Subject and product information are coded, so that analysis of results is performed "blind" according to the standard.

[0115] The microcomputer includes in its memory means, or via network access to a shared application, a program for processing images from the multispectral camera to carry out spectral unmixing of an image, preferably via the SUnSAL method, and to isolate the components attributable to erythema in said images.

[0116] A classification program allows the response curve to be determined from the results of the image processing program (420, 421 figures 4 to 6 ) corresponding to each of these acquisitions.

[0117] Following the completion of this treatment, the micro-program in question or a complementary display program makes it possible to display an image (710) in false colors showing the intensity of the erythemal component on the different spots of the acquired areas and to compare the results obtained on the protected area (711) and on the unprotected area (712).

[0118] The display of this image makes it possible, among other things, to facilitate the determination of the DEMn and DEMp by visually reinforcing the erythemal component. Image processing not only makes it possible to restore information on the intensity of the erythema, by displaying in false colors, but also spatial information on the extent of the erythemal zone within each spot, which facilitates the determination of the areas corresponding to the DEMn and DEMp.

[0119] Thus, according to a first embodiment, the analyst points to the area (spot) corresponding according to him to the DEMn, then the area (spot) corresponding to the DEMp.

[0120] Still according to this first embodiment, after validation, the program extracts from the data files corresponding to the UV doses associated with these zones and calculates the FPSi using the response curves as presented above with reference to the figure 5 , before storing the resulting value in a result file.

[0121] According to another embodiment, the FPSi is determined from a threshold or a band as illustrated with reference to the figure 6. For this purpose the threshold or band, or more advantageously a plurality of thresholds or a plurality of bands are stored in the memory means of the computing device (701) associated for example with skin color characteristics. The selection of the appropriate threshold or band is carried out on the basis of the metadata transmitted with the images.

[0122] According to another embodiment, the response curves of the protected part and the unprotected part are approximated by sigmoids and the FPSi of the product is determined by a program aimed at minimizing the mean square deviation between the response curve of the protected area and the response curve of the unprotected area affected by a coefficient k, k being the FPSi. This embodiment makes it possible to automate the determination of the FPSi by analyzing the images.

[0123] Some or all of these embodiments are advantageously combined and thus provide for a given acquisition, a plurality of FPSi values. According to this combined embodiment example, the FPSi value obtained is considered valid if the plurality of values ​​corresponding to the different methods and included in an interval, for example + / - 10% centered on the average value of the plurality.

[0124] After processing all valid results corresponding to the tested product and stored in the results file are used to determine the SPF of the product.

[0125] The method implemented by the system which is the subject of the invention is improved if, in addition to the areas exposed to irritation, whatever the origin of this irritation (exposure to UV or to a product in topical application), the acquisition of an area of ​​healthy skin, close to the test area, is carried out. Thus, this acquisition is advantageously used to determine the spectral components of the erythema, as explained with reference to the figure 2 , or to check the validity of the response curves as explained with reference to the figure 3 .

[0126] Advantageously, the acquisition carried out on healthy skin is associated with a target. This embodiment allows, among other things, to determine the color of the subject's skin by image processing.

[0127] The embodiments set out above are mainly set out in the context of determining the sun protection factor of a photoprotective product, however all or part of the principles and embodiments set out can be used for the characterization of erythemal areas in the context of other skin irritation tests, implementing comparisons of erythemal areas or comparisons between irritated areas and healthy areas.

[0128] The possibility of automatic characterization of the areas subjected to the irritation test by means of the device and the method which are the subject of the invention makes it possible for certain applications to process a large number of tests quickly and reliably.

[0129] Thus, the method and system which are the subject of the invention are advantageously used for systematic tests within the framework of screening the ingredients likely to enter into the composition of a topical application product, with regard to the irritant power of these ingredients or their protective effect. Example

[0130] In order to test the validity of the method, a study to determine the FPSi of a product, referenced P2 in the ISO 24444:2010 standard, is carried out on 10 subjects.

[0131] Each subject is thus exposed to UV on 2 areas on the back, first area without photo-protective agent, a second with the photo-protective product P2. This product is composed of 3 phases, phase 1: lanolin 4.5%; Theobroma cocoa butter 2.0%; glyceryl monostearate 3.0%; stearic acid 2.0%; ethylhexyldimethyl PABA (CAS 21245-02-3) 7.0%; benzophenone-3 (CAS 131-57-7) 3.0%; phase 2: water 71.6%; sorbitol (70% liquid) 5.0%; triethanolamine 1.0%; methylparaben 0.3%; propylparaben 0.1%; phase 3: benzyl alcohol 0.5%. The data correspond to mass fractions (%).

[0132] Each exposure zone consists of 6 spots with a dose progression of 1.25 (in accordance with the standard). For each subject, zone 2 (with photoprotective product) was exposed to doses 14 times greater than zone 1. The exposure system used is UV SOLAR LIGHT MULTIPORT BERGER (model 601 of 300 W) whose spectrum complies with the standard.

[0133] 24 hours after exposure, each area of ​​each subject is subject to multispectral acquisition using the device that is the subject of the invention, comprising a multispectral camera with 19 wavelengths between 365 nm and 970 nm, and the value of the FPSi corresponding to each of the subjects is determined by comparing the response curves of the protected part and the unprotected part after spectral demixing and extraction of the erythemal component.

[0134] The distribution of camera acquisition wavelengths is given in Table 1 below. Band number Band name Wavelength [nm] 1 Grapes 365 2 Violet 405 3 Indigo 435 4 Blue 450 5 Blue 470 6 Cyan 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

[0135] The corresponding results of the FPSi calculation are given for each subject in Table 2 below: Subject No. 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

[0136] The arithmetic mean of the SPFs provides the product's SPF value, in this case an SPF of 16.06.

[0137] This value of 16.06 corresponds to the reference value of the standard for the P2 product which is 16.1; which standard indicates in its Annex C an average FPS of 16.1 and between 13.7 and 18.5.

[0138] This example shows the relevance of the method which is the subject of the invention for the automatic determination of an FPSi.

Claims

1. Method for determining the individual sun protection factor, SPFi, of a photoprotective product from 2 images each of them comprising a plurality of spots obtained by exposing a skin to increasing UV doses, one of the series being obtained on a skin without protection and the other on the skin protected by the product, the two series causing an erythemal character to appear on at least one of the spots, comprising the steps of: a. acquiring a multispectral image (711, 712) at discrete wavelengths ranging from 365 nanometers to 970 nanometers, of each of the series; b. carrying out for each of the images a spectral demixing so as to isolate in said images the spectral components (212) corresponding to erythema, said spectral components (212) corresponding to the modification, attributable to erythema, of the spectrum of the skin compared to the spectrum of the healthy skin, so as to obtain for each spot of each of the series the intensity (402) of said spectral components (212) corresponding to erythema; c. determining from the result of step b) a response curve (420, 421) giving the intensity of said spectral components (212) corresponding to erythema according to the UV dose for each of the series; d. comparing the response curves (420, 421) thus obtained in order to determine the SPFi.

2. Method according to claim 1, wherein step d) comprises approximating the response curves (420, 421) by sigmoids fn, fp and where the SPFi is the k value minimizing the mean square deviation between the functions fn, fp over a range from the lowest UV dose to the highest UV dose of exposure of the compared zones.

3. Method according to claim 1, wherein step a) comprises acquiring a zone of healthy skin (without erythema) and that step b) of spectral demixing uses this acquisition for determining said spectral components (212) corresponding to erythema.

4. Method according to claim 1, wherein the multispectral image comprises at least 8 primaries per pixel corresponding to wavelengths covering 400 nm to 700 nm.

5. System for implementing the method according to claim 1, comprising a multispectral camera (701) for acquiring a multispectral image (711, 712) at discrete wavelengths ranging from 365 nanometers to 970 nanometers of exposed zones of a skin, a computer processing device (702) comprising memory, display, input and calculation means, capable of retrieving the images acquired by the multispectral camera, characterized in that it comprises an image processing program capable of performing spectral demixing of the acquired images and isolating from said images the spectral components (212) corresponding to erythema, said spectral components (212) corresponding to the modification, attributable to erythema, of the spectrum of the skin compared to the spectrum of the healthy skin, and a program for comparing the images based on the intensity of said spectral components (212) corresponding to erythema of each image.

6. System according to claim 5, wherein the images acquired by the multispectral camera (701) and transmitted to the computer system (702) comprise metadata relating to the conditions for creating erythemal zones and to the conditions for acquiring the images.

7. System according to claim 5, wherein the memory means of the computer device comprise a threshold (620) or a band (630) of erythemal intensity depending on the UV dose (401), and wherein the comparison program determines the SPFi by the intersections of this band (630) or of this threshold (620) with the response curves (420, 421) of each compared zone.

8. Use of the system according to claim 5 for implementing the method according to claim 2, for carrying out screening in view of selecting the ingredients constituting a product intended for a topical application.