WO2013098512A1

WO2013098512A1 - Method and device for detecting and quantifying cutaneous signs on an area of skin

Info

Publication number: WO2013098512A1
Application number: PCT/FR2012/053028
Authority: WO
Inventors: Sergio Dos Santos; Benjamin Boulay
Original assignee: Chanel Parfums Beaute
Priority date: 2011-12-26
Filing date: 2012-12-20
Publication date: 2013-07-04
Also published as: FR2984721B1; JP2015500722A; KR20140112046A; FR2984721A1

Abstract

Method for detecting and quantifying cutaneous signs on an area of skin, according to which: * for a plurality of predetermined classes of degrees of skin colour intensity and a plurality of types of cutaneous signs, a plurality of selected wavelengths are identified such that, for each combination of a type of skin sign and class of degree of skin colour intensity, one of the wavelengths can be used for detecting by means of contrast such a type of sign on the skin of a person having a degree of skin colour intensity belonging to this class; * images are taken in white light; the degree of skin colour intensity and the corresponding class are automatically identified from said images; according to this class of degree of skin colour intensity, images are automatically captured at the wavelengths that correspond to the combination of this class of degree of skin colour intensity with a type of sign that it is sought to detect, then at least this captured image is processed, so as to detect signs of said type that it is sought to detect.

Description

METHOD AND DEVICE FOR DETECTION AND

QUANTIFICATION OF SKIN SIGNS ON A SKIN AREA

The invention relates to a method and a device for detecting and quantifying, automatically (or at least semi-automatically), various cutaneous signs on a skin zone, particularly on an area of the face (or even the neck); these signs can include dark circles, wrinkles, tasks, etc.). These types of signs can be quantified by "grades".

In this document, various words will be used with the following meanings:

- Cutaneous sign: modification of the area of skin considered, likely to deserve a cosmetic or therapeutic treatment; these skin signs may be signs related to skin aging, dryness, fatigue or changes in skin color. o "signs of skin aging" means any changes in the surface appearance of the skin due to aging such as, for example, wrinkles and fine lines, stains, all signs of sagging skin, changes in skin thickness, lack of elasticity and / or firmness of the skin, dull and lackluster skin.

o "Signs of skin dryness" means any changes in the surface appearance of the skin due in particular to changes in the water content and its distribution within the Stratum Corneum, such as the dull, rough and squamous, non-silky, reddish and / or scaly, as well as loss of flexibility and change in skin thickness. Signs of dry skin include drought-related sensations such as itching, tingling, and / or tugging, which can result in the appearance of real pathologies such as, for example, hypersensitivity, atopic dermatitis or winter xerosis,

o "Fatigue signs" and "skin color changes" means any changes in the surface appearance of the skin due to fatigue or poor blood circulation, such as dark circles. or bags under the eyes, as well as any changes in skin color, including an impression of aggravation of the signs of aging of the face resulting from exposure to different lifestyles (exposure to the sun, sleep deprivation, stress, jet lag ...),

Grade of a cutaneous (or cosmetological) sign (or criterion): degree of severity / severity of this sign or criterion; the value of this degree is also called the descriptor of this type of sign,

Degree of intensity of skin color: parameter quantifying the more or less light or darkness of a skin zone, independently of any indication of color; the degrees can be determined on color images (three RGB components) by a statistical classification analysis, preferably a Principal Component Analysis (PCA), or even by an Independent Component Analysis (ACI), a Factoring in Non-Negative Matrices (FMNN) or any other similar classification method known to those skilled in the art; dividing the various possible values of skin color intensity levels into a plurality of at least two classes.

Monochrome or monochromatic image: image associated with a given wavelength; more precisely, a monochrome image is taken (without particular filtering) under a lighting at a given frequency. Such a monochrome image can be captured differentially, without this being mandatory.

- White illumination: lighting on most of the visible spectrum (ie about 400 nm to 700 nm), whatever the way to obtain this lighting: white LEDs, RGB LEDs, halogen light, fluorescent lamps incandescent, fluorescent tubes, fluorescent lamps, ...).

- Morphological operation: operation associated with a type of cutaneous sign to detect and locate on an image the presence of this type of cutaneous sign.

- Image rate: image resulting from the ratio between two filtered images of the same area (in each pixel, the image rate replaces the value of this pixel by the ratio between the average pixel values of a zone directly surrounding the pixel , and the average of the values of the pixels of the neighborhood of this zone).

US-2009/0201365 (Fukuoka et al) discloses a system for diagnosing the condition of a skin area and developing skin treatment tips. This system comprises a set of data collection (or even a plurality of such sets) and a set of analysis of these data, which cooperate via means of communication. The data collection assembly includes a very high resolution digital image capture device, a high compression compression device, and a display device; the data analysis assembly comprises a data analysis device and a compression set with a high compression ratio and data storage means. The data collection assembly takes an image of the entire face of a subject, and the compression means creates a compressed image, the data analysis device analyzes the condition of the skin from this compressed image received through the means of communication and creates visual information as a diagnostic result; this visual information is compressed and then sent to the data collection assembly for viewing by the display device. The data analysis set is independent of the data capture set and may be remote from the capture set. The capture of the data is done by a simple digital camera, under predetermined conditions, which can be done by an operator without special training; as for the analysis operations, they are done by reference to very rich databases, on a multitude of criteria.

Such a system is complex and involves the development of very rich reference bases, intended to separate various possible items of analysis.

Also known from US-2004/0218810 are methods and systems for computer analysis of images of the skin. According to this document, a digital image of a face area illuminated by white light is taken by means of an RGB sensor and the images R (Red), V (Green) and B (Blue) are processed. image thus captured by various analyzes (skin color, tasks, pores, etc.) allowing to deduce various data allowing, if necessary, to evaluate an evolution over time. The images are taken using a device involving precise positioning of the face, providing in practice a color reference near the face. But this system may seem uncomfortable, while the analyzes are complex, analyzes often have to be performed on several R, G or B images of the same area, with in practice HSV conversions, with a precision that often depends from comparisons to references.

Similarly, according to the document WO-94/16622, a method of acquiring images of an illuminated zone in white light or in UV light for the detection of cutaneous lesions is known. An essential feature of this method is the selection of light bands in the secondary beam, reflected by the illuminated area in white light; this selection is essential for the final result of detection and classification of cutaneous lesions. But this selection implies the application to the beam secondary filtering, for example by means of a filter wheel which involves successively treating the light bands of interest, or by means of a beam splitter which allows simultaneous processing of these bands, to price, however, of a significant loss of signal-to-noise ratio on each band.

Document US-2009/240653 discloses a method for acquiring images of a zone illuminated with white light making it possible to determine the phototype of the skin of a volunteer, among the six possible phototypes according to the Fitzpatrick classification. ; this method involves modeling the histograms of the bands R, G and B of the image of a volunteer's area by Gaussian distributions and it is from the means and standard deviations of each distribution that the Fitzpatrick phototype is deduced by a tree algorithm. The subject of the invention is a system for analyzing skin zones that is simpler and more compact, while being able to involve means of calculation of moderate size, requiring, for example, nothing more than a PC of reasonable performance.

It proposes for this purpose a method for detecting and quantifying cutaneous signs on a skin zone, according to which:

for a plurality of at least two classes of predetermined skin color intensity levels and a plurality of predetermined types of cutaneous signs, a plurality of selected wavelengths are identified beforehand so that, at each combining any type of skin sign within this plurality of types and any class of skin color intensity level within said plurality of classes, at least one of the wavelengths of the plurality of wavelengths allows the contrast detection of such a type of sign on the skin of a person having a degree of skin color intensity located within said class,

* we take under white lighting at least one image of a selected area of the face of a subject, ^* Automatically is identified from this the degree of skin color intensity corresponding to the person and identifies the class to which this degree,

^* Is selected, automatically as a function of this class of degree of skin color intensity, one (s) of the wavelengths of said plurality which allow (s) to this degree of skin color intensity detection of a type of sign that one seeks to detect on the chosen zone of the face,

at least one monochromatic image is acquired at least this (these) wavelength (s), which is processed using an algorithm and parameters mainly chosen according to the degree class of intensity of skin color, so as to detect and graft signs of said type that one seeks to detect.

Thus, the invention takes advantage of the observation that it has been possible to make an analysis of cutaneous signs by means of algorithms and parameters in limited numbers, from monochromatic (or monochromatic) images - see the definition below. above) corresponding to carefully chosen frequencies, taking into account mainly the intensity of the skin color. The acquisition of some images of an area illuminated by a few well-chosen particular wavelengths makes it possible to detect a large part of the cosmetological signs and this, for all the degrees of intensity of existing skin color; in this context, it is sufficient to separate the degrees of intensity of skin color in a very small number, freely chosen and can be as low as two, to be able to detect a large amount of cosmetological signs in a wide variety of possible volunteers.

The invention thus proves to be particularly simple since, in particular, it makes it possible to freely choose a very limited number of degrees of skin color, without having to take into account, in particular, the six Fitzpatrick phototypes as in document US-2009/240653 and therefore without having to implement an equivalent number of algorithms and groups of parameters while allowing, unlike WO - 94/16622, a more complete use of the secondary beam, that is to say reflected by the zone In progress inspection since the invention does not require filtering or separation of this secondary beam. It should be noted that part of the advantages of the invention results from the illumination of the area of interest at selected wavelengths (by being able to use the entire reflected signal) whereas the documents cited imply in practice white light lighting using only a portion of the reflected beam.

Advantageously, this plurality of wavelengths is preferably between two and five for all classes of degrees of intensity of skin color; it has indeed appeared possible to detect several types of cutaneous signs, for all the degrees of intensity of skin color, with barely two wavelengths, preferably chosen equal to the order of 500 nm ( plus or minus 20 nm) and 570 nm (plus or minus 20 nm). However, preferably, the plurality of wavelengths for all classes of skin color intensity levels further comprises a wavelength of around 620 nm (plus or minus 20 nm). ). Thus, the number of wavelengths for all the classes is advantageously at most equal to three.

Preferably, the plurality of types of cutaneous signs that the method makes it possible to detect and quantify include, in particular, wrinkles, dark circles and pigment irregularities.

Advantageously, for at least one of the degrees of intensity of the skin, two wavelengths are chosen for the same type of cutaneous sign (or even several types of cutaneous signs). In a preferred configuration, the plurality of wavelengths is comprised of three wavelengths and the plurality of classes of skin color intensity levels are comprised of two classes, namely the skin class and the class dark skin. Despite this low number of wavelengths and classes of degrees of skin color intensity, it is possible to detect and quantify a wide variety of cutaneous skin signs, including wrinkles, dark circles, spots, and so on. More particularly advantageously, in a simple but effective configuration, one affects at most two wavelengths for types of cutaneous signs for fair skin, and at most two wavelengths for the types of cutaneous signs for dark skin.

Advantageously, the capture of each monochrome image corresponding to a given wavelength comprises the capture of two successive images of the same skin zone taken with and without illumination at the corresponding frequency, and the elaboration of the monochrome image by difference of these two images; this eliminates the influence of ambient lighting, especially its possible fluctuations during the day or its possible spatial fluctuations. More preferably, a plurality of images for the same wavelength are captured at increasing levels of illumination power, and the image obtained with the highest power level is selected without being saturated.

The invention further proposes, for the implementation of the method, a device for detecting and quantifying cutaneous signs on skin zones, comprising:

^* An illumination device adapted to illuminate an area of skin placed in a given location, under various illumination schemes, optionally,

^* An image capture device adapted to capture a monochromatic image of an area of skin placed in that particular location,

^* a treatment device,

^* Interfaces between this capture device, illuminating device and processing device, such as the processing device processes images captured by the image capture device in the lighting determined by the illumination device,

this treatment device:

containing identification data, for a plurality of at least two classes of predetermined skin color intensity levels and a plurality of skin sign types, of a plurality of selected wavelengths so that, at each combination of any type of skin cutaneous sign at within the plurality of types and any class of skin color intensity degree within the plurality of classes, one of the wavelengths of the plurality of wavelengths allows the detection by contrast such a type of sign on the skin of a person having a degree of skin color intensity within that class,

being designed so as to cooperate with the capture device and the illumination device for capturing images of the chosen skin zone in white light and at said wavelengths of said plurality, the illumination regimes allowing a capture in white light and capture at each of the wavelengths of this plurality, and

being designed so as to automatically identify in at least one image captured by the white-light image-capturing device the degree of skin-color intensity corresponding to the person concerned and the class to which this degree of intensity belongs of skin color, to select, automatically according to this class of degree of intensity of skin color, that (s) of the wavelengths which allows (tent), for this class of degree of intensity of color of skin, the detection of a type of cutaneous sign that one seeks to detect, to automatically capture at least one image at least this selected wavelength and to automatically process at least this captured image at this wavelength selected, using an algorithm and parameters mainly selected according to the class of degree of intensity of skin color, so as to detect signs of the type that is to be detected.

Objects, features and advantages of the invention will become apparent from the following, given with reference to the accompanying drawings, in which: FIG. 1 is a schematic diagram of the method of the invention; FIG. 2 is a block diagram of a system adapted to the implementation of the method of the invention,

FIG. 3 is a sectional view of the pair formed by the camera and the device for illuminating the face of the subject,

FIG. 4 is a graph of an example of an illumination sequence at a given length for the differential recording of images,

FIG. 5 is a graph correlating two image components captured for a wide variety of volunteers,

FIG. 6 is a graph grouping in another way the results of FIG. 5,

FIG. 7 is an example of a processed image in order to detect and quantify pores,

FIG. 8 is an example of an image processed so as to detect and quantify dark circles under one eye,

FIG. 9 is a schematic diagram of the construction of a "rate" image used in the detection of pores, scales, spots and telangiectasias,

FIG. 10 is an example of highlighting pores by the construction of the "rate" image,

FIG. 11 is an example of a pore detection algorithm,

FIG. 12 is an example of a dander detection algorithm,

Fig. 13 is an example of spot detection for two classes of skin color intensity levels, and

FIG. 14 is an example of separation of spots / telangiectasia by Principal Component Analysis (PCA), the component 1 being plotted as the abscissa at the bottom, the first two components being reported in abscissa at the top, and component 2 is plotted on the y-axis.

As is apparent from Figure 1, the method of the invention mainly comprises the following steps, for detecting and quantifying cutaneous signs on a skin area.

In advance (preliminary parallel steps 1 and 2), a plurality of classes (at least two) of predetermined degrees of skin color intensity (denoted Di) and a plurality of types of cutaneous signs (denoted Si) are identified. It is then identified (preliminary step 3), according to these classes of degrees Di and these types of signs Si, a plurality of wavelengths Ai chosen so that, at each combination of any type of cutaneous sign and of any class of degree of skin color intensity, one of the wavelengths of the plurality allows the contrast detection of such a type of sign on the skin of a person having a degree of intensity of skin color entering the class considered. The identification of these wavelengths also includes possible special imaging conditions at such wavelengths (see for example below, concerning the conditions of the illumination intensities, and / or the taking of differential images). In other words, for each of the classes, acquisition conditions are identified for obtaining images enabling the detection and quantification of the various types of cutaneous signs that one wishes to be able to detect and to quantify. In addition, processing conditions (in practice algorithms and possible parameters) of such images are determined for effecting such detection and quantification, in practice automatically.

As an example, as will be discussed later, it is sufficient to distinguish two classes of intensity levels of skin color, namely "light skin" and "dark skin" (for example D1 noted and D2), and define for each of the classes algorithms for detecting and quantifying the various types of cutaneous signs considered. The choice of the types of cutaneous signs, and their number, depends on the needs; for example, it is advantageous to distinguish:

- the wrinkles, possibly distinguishing, according to their locations

- the naso-genial fold (in English "naso-labial fold"),

- forehead wrinkles,

wrinkles called "wrinkles of the Lion", located between the eyebrows (in English "frown lines"),

- Wrinkles called "crow's feet wrinkles", located at the corner of the eyes (in English "crow's feet"),

- wrinkles under the eyes,

- the wrinkles of the corner of the mouth,

- wrinkles of the contour of the lips,

- stains, possibly distinguishing between freckles and age spots and

- pores and scales

- Telangiectasia ("telangiectasia" in English), that is to say the small dilations of the superficial vessels.

The wavelengths are chosen so as to allow, for each of the classes of degree of intensity of skin color, to identify at least some of the types of cutaneous signs that one wishes to be able to detect and quantify. As will appear below, it is possible to retain only three wavelengths for all classes of degree of intensity of skin color, to be able to detect and quantify signs of the aforementioned types, both in the light skin only in dark skin; in practice it seems useless to implement more than 6 (or even 5) different lengths (and thus to determine more than 6 or even 5 different processing algorithms for the various classes). Similarly it seems useless to implement more than 4 (or even 3) classes of degree of intensity of skin color.

As indicated above, taking a monochrome image, corresponding to a given frequency, corresponds to a seizure panchromatic (in the visible range) of an image of an object receiving a lighting limited to this specific frequency. In other words, the monochrome image capture is to illuminate an area with monochromatic illumination, but to capture all radiation from the observed area, in response to this monochrome illumination, without specific filtering other than that which may exist in the image capture device. This is in practice the same image capture device that captures the image in white light (that is, the image illuminated by white illumination) and the monochrome (or monochromatic) images.

The various correlations between the various wavelengths and the various "type of sign" / "intensity degree class" pairs may have been validated and recorded beforehand (for example before the actual commissioning of the setting device during a step that can be described as "sampling"), so that, in practice, the implementation of the method, for a given person, can be limited to the following steps.

We begin (step 4) by taking a white-light image of a selected area of a subject's face (or even the face as a whole). The triggering of this step can continue, in an automated way, by the other steps

From this image is identified (step 5) the degree of skin color intensity corresponding to the person concerned. Then, depending on the degree of skin color intensity, that of the classes in which the degree of skin color intensity is selected (this is done in practice automatically from the data collected during the preliminary stages. ).

We acquire (step 6) the images corresponding, for this class, to the types of signs that we want to detect, according to the aforementioned acquisition parameters, previously defined for this class (illumination wavelengths in particular,. ..). If the number of types of cutaneous signs to be detected and quantified is small, it is possible that the image acquisition does not involve all the selected wavelengths: in fact, it is preferable to acquire only the images that we need for detection and the quantification of the types of cutaneous signs. The selection of wavelengths is automatic from the identification of the class to which the skin under examination belongs; the triggering of this acquisition can also be automatic. The implementation of the method can be done automatically from the moment an operator triggers the image in white light; it can also be done in several times, for example in a first phase since the triggering of such a white light image pick up to the selection of wavelengths, then a second phase of taking the image (or images) under illumination at this (these) selected wavelength (s) until the end of the treatment (for example until the results of this treatment are displayed. that the treatment is limited to detecting some of the possible cutaneous signs, this choice can be entered by the operator before the triggering of the white-light image, or before the triggering of the second aforementioned phase.

It should be noted that, although the sampling has been done for a given number of types of cutaneous signs, it is possible, during the implementation of the method, to search for a smaller number of types of signs, according to the specific needs expressed by each client (the sampling of steps 1 to 3 consists in anticipating all the needs that one wishes to be able to satisfy later, without obligating to use all the combinations).

These acquired images are then processed in the selected conditions (step 7), using at least one algorithm and parameters previously essentially chosen according to the class of degree of intensity of skin color, so as to detect types of signs that one seeks to detect. This treatment is like detecting the signs and grading their intensity.

Finally, in any appropriate form (graphically or in the form of a table, in particular, as needed), the detection and gradation results are displayed for the types of signs selected (step 8). It is understood that such a display step can be omitted, especially if, for example, the process is completed by a step of automatic prescription of products adapted to the treatment of the type of cutaneous sign considered, with the grade noted. Thus, more generally, the results are stored and / or displayed.

Figure 2 schematically shows a system adapted to the implementation of such a method.

It mainly comprises:

an image capture (or capture) device, such as a digital camera 11, designed to capture all the radiation from a location where an area to be observed is to be placed (there is no filter favoring one or another frequency within the radiation reaching this capture device),

an illumination device 12 comprising sources in white light and at the various selected wavelengths,

interfaces 13 and 14 associated with this camera and this illumination device,

an interface device 15 with a computing device which can be a simple PC,

a computing device, such as a PC 16, comprising modules (not shown) adapted to carry out the operations of the method, automatically,

a power source 17.

It should be noted that such a device is compact and is easy to move, if necessary.

Fig. 3 shows a preferred configuration of the image capture device and the illumination device. This illumination device 12 comprises a concave surface intended to be turned towards a subject whose skin condition is to be diagnosed, preferably materialized by a concave diffuser screen 12B, for example in "curved plexiglass®". Behind this screen are distributed sources generally schematically under the reference 12A, which are monochrome, each emitting at one of the chosen wavelengths, or sources of white light; the sources emitting at each given wavelength are advantageously regularly distributed among the others, so that, when the sources corresponding to only one of these wavelengths are fed, they provide a uniform and uniform illumination on the skin zone to characterize (it is the same for the sources of white light).

Advantageously, the illumination system 12 contains LEDs respectively corresponding to each of the selected wavelengths, also uniformly distributed, and LEDs of white light, also uniformly distributed, with appropriate power supplies for the various image acquisition (there may be the possibility of implementing various intensity levels for the various sources).

The image pickup device 11 is preferably disposed in the center of the illumination surface formed by the diffusion screen, crossing the latter.

It will be understood that the subject whose skin condition is to be diagnosed is placed in a position centered with respect to the curved surface of the device of FIG. 3; we then proceed to successive captures of selected images, first in white light, then at useful wavelengths. Depending on the needs, the ambient lighting (apart from the illumination that the illumination device 12 can cause) can be modified, or not, at the time of the image captures.

Advantageously, the image captures result from the taking of two successive images, one of which is taken with the ambient lighting, and the other is taken with the same ambient lighting but adding lighting to the frequency / wavelength to which the image to be captured must correspond; this monochrome lighting must be powerful enough to cause a difference between the two successive raw images, but not too powerful not to risk causing saturation within the image capture device. By difference between the successive images, one identifies the captured image at the frequency in question. The foregoing may furthermore apply to white-light imaging, if desired (it may however be sufficient to take single images in white light and differential images in monochrome light).

FIG. 4 represents, by way of example, an illumination sequence during an imaging session (both in white light and for the selected wavelengths).

This graph represents a sequence of taking three successive images, for the same lighting mode; a first cycle consists of taking (at the instant designated by "Shot 1") a first image by applying to the LED diodes associated with this lighting mode a low level of excitation and then an image without excitation of these LEDs (at the instant designated by "shot dif"); a second cycle is then triggered differing from the first cycle in that the excitation level of the lighting mode in question is more important (here there are also two image captures, at times designated by "Shot2" then " Shot dif "), and a third cycle is here then triggered with a maximum level of excitation for this mode (here again with capture of two images, at times" Shot3 "and" Shot dif "). The "Shot dif" instants are designated in each of the cycles by the same name to signify that, in principle, the corresponding images are identical, being taken with the same ambient lighting on the same area of skin.

It will be understood that the duration of the excitation pulses applied to the diodes associated with each illumination mode (white light or wavelength), as well as the intervals between the image pickups, is arbitrary; for example, the cycles follow each other with intervals of less than one second. In fact the variations of luminosity and color felt by the subject can contribute to the relaxation and thus to improve the relevance of the diagnoses that will be made on the basis of the images thus captured. The image which is then conserved is preferably that corresponding to the maximum intensity of the diodes (in particular in the case of the taking of monochrome images) without having saturation, but the cycles prerequisites will have minimized any risk of clenching the subject at the moment of maximum illumination.

Such an approach makes it possible to overcome the difficulties due to the fact that:

- the ambient lighting can not be controlled,

the pose by the subject must not be too long for the subject to cause him unwanted movements.

The results of the image processing are advantageously visualized, on a screen or printed: these results preferably include the identification of the types of signs explored, with an indication of the level of their presence (frequency and / or size, depending on the parameterization of the algorithms used). By way of example, an image acquisition in white light is then carried out, then:

* we distinguish whether it is a light or dark skin (in the case of only two classes of degree of intensity of color),

^* we acquire images according to this distinction, * we launch algorithms to detect the different cosmetological criteria in these images,

^* for each of the algorithms, one stores the data necessary for the classification (which allow a gradation for each criterion of the face considered)

* we save the results (especially visual) detection in a predefined directory.

For example, we write the results in a file that can be only a text file such as: Name of the subject and: C002 Skin Claire

Nasolian furrow: Grade 3

Wrinkles Front: Grade 4

Lion Wrinkles: Grade 3

Wrinkles of the crow's feet: Grade 2 Eye Wrinkles: Grade 5 Wrinkles from the corner of the mouth: Grade 3

Lip contour wrinkles: Grade 2

Front spots: Grade 1

Cheek spots: Grade 1

Telangiectasia: Grade 2

Pores: Grade 1

Squames: Grade 1

Dark circles: Grade 2

By way of example, grade 1 corresponds to a low or zero presence while level 5 corresponds to a maximum.

The stored results may also include at least one of the images taken into consideration.

Degree of intensity of skin color

It may be noted that the invention does not involve traditional notions of ethnicity, such as Caucasian, Asian, African, Hispanic or Indian; indeed, it only involves, at the time of treatment selections, the notion of degree of intensity of skin color.

We performed classification analyzes on a panel of photographs. It emerged that, regardless of ethnicity, globally homogeneous groups could be identified when the intensity of skin color was included as a characteristic.

This could be explained, and justified, with different clustering methods. In the example given here in FIG. 5, which graphically represents an analysis made on a photographic database of 470 photos of volunteers, we used the analysis known as "Principal Component Analysis" (ACP for short). ). It should be noted that we could have used other classification methods such as, for example, those known under the names of "Non-negative Matrix Factorization (NMF)" or "Independent Component Analysis" (ICA). ) ".

For each of these images, the average values of the RGB components (ie, red, green and blue) of the face were measured and a PCA applied to this average component. The results of the ACP are presented in Figure 5 which represents all the points corresponding to the volunteers in the 470 photos, placed in the frame formed by the first two main components of the ACP. These points were arbitrarily separated into 5 regions (symbolized by 5 rectangles) by setting 4 thresholds on the values of the first principal component. These results clearly show that the first major component is associated with the light or dark nature of the skin, regardless of ethnicity. For example, Hispanic volunteers are scattered between the first rectangle (with an average value of the first component of -0.3) and the second to last (with an average value of 0.25).

The results of the PCA show very clearly that the first main component explains almost all the skin color variation between volunteers: 98.7% of this variation can be modeled by a variation of the value of this component. For example, the volunteer located at the bottom right of Figure 5 has a value of this component greater than 0.5, while the lower left has a value of the order of -0.2. The variation between -0.4 and 0.6 of the value of the first main component is therefore able to represent skin color variations of all ethnicities. It is therefore possible to identify the degree of skin color intensity from this first component.

It is thus justified to base the choices on the skin color (more precisely on the degree of intensity of skin color) rather than on the notion of ethnicity, which allows a great simplification of the treatments, since it suffices advantageously to distinguish two classes of degrees of intensity of skin color. The same type of results can be obtained with other classification methods (Independent Component Analysis, known as ACI (or ICA for Independent Component Analysis), or factorization in non-negative matrices, known in English under the acronym NMF for: "Non-negative Matrix factorization"). In addition, it should be noted that depending on the method used, it is possible to have a separation in as many classes of skin color intensity as desired. Choice of degree of intensity of skin color

Figure 6 shows the histogram of the values of the first major component for the 470 volunteers. We note two components in this histogram, highlighted by the curved lines defining the envelope. The characteristic threshold separating these two components can be found automatically by a simple histogram binarization algorithm (such as the Otsu algorithm): a threshold equal to S _{0 is found} .

The ACP allows to find a threshold with a component. We could find n thresholds to have n + 1 classes according to the clustering method used.

Thus, if the value of the first principal component for a volunteer is greater than SQ, then this volunteer will be considered to have a dark skin and she will be considered light-skinned if not. This makes it easy to distribute different subjects in only two classes of degree of color intensity.

Degree of intensity of skin color for a new subject

Suppose a new subject image is provided, it remains to classify it in one of the two skin groups.

To do this, Principal Component Analysis is nothing more than a basic change between two benchmarks, the first of which is (R, G, B) (measured) and the second (x, y, z). ) which minimizes the variance of the final set of points. To classify the new image, it is therefore sufficient to compare its coordinate x ₀ in the new axis system with the threshold determined for the test group (see above). For this, we measure the average values <R>, <G>, and <B> in the new image. Once these <R>, <G> and <B> values are obtained, this means that the new image is seen as a point in the repository (R, G, B). It is therefore necessary to change base and this is done naturally in matrix calculation with the base change matrix obtained for the test set. The coefficients of the transformation are represented in the form of a matrix, each column of which corresponds to a component. In a manner known per se, these columns are classified according to the most significant principal component to the least significant component. We know how to deduce the coordinates of the point in the new repository. The coordinate point (xo, yo, zo) thus obtained is the representative point of the new person in the principal component database and it is the value x ₀ that can be compared with the threshold So to decide in which class the new volunteer: if it is below the threshold, it will be a clear skin, and if it is above it will be a dark skin.

After acquiring the necessary images, under the predetermined conditions indicated above, they are processed so that they can search and quantify at least one type of cutaneous sign, using appropriate algorithms.

Detection algorithms

Various examples are given below as to how one can, by the appropriate choice of algorithms and / or their parameters, characterize the various types of cutaneous signs. Figures 7 and 8 show by way of example a representation of pores, or a representation of rings. at. Image "rate"

According to an aspect of the original invention per se, we introduce here an innovative transformation of a starting image into an image which is the ratio of two filtered images, which we call "image rate". This image rate advantageously serves as a basis for the detection of pores and dander, but can also be useful, in particular, for detecting stains.

A rate image is a filtered version of the image, which respects the local contrast between a pixel (or pixel area) and its neighborhood. It is estimated as the ratio between the average of the pixels of an area and the average of the pixels of its neighborhood. If a zone is darker than its neighborhood, the average of the pixels of this same zone is less than the average of the pixels of its neighborhood, and therefore the image rate will have a value lower than 1 in this zone. The image rate is the ratio between two filtered versions of the original image. The first filter is a convolution of the image with a "medium" filter (3x3 in Figure 9). The second filtering is a convolution with a "medium" filter, in which the central pixels are separated (7x7 filter in the figure, with 3x3 central pixels set to zero). Figure 10 shows the transformation of an image (left) into a rate image (right), and the effect this has on pores, stains and fine lines.

This local transformation of an image makes it possible to highlight certain dermatological criteria, such as the pores in the example above. Indeed, a dark value in the image rate implies that the starting area in the initial image was darker than its neighborhood. However, it is recommended that the size of the first filtering be representative of the size of a pore in an image (when the criterion is the pore). The image rate created is uniform, which advantageously overcomes the problem of lighting. This image highlights moles, spots and wrinkles (but can not in any way mitigate the problem of blurred images). b. Pores / Dander

Pore and dandruff detection algorithms follow the same principle (pores are often small dark areas on a light background, while dander are often small, light areas on a dark background). Here we present in detail the pore detection algorithm, and only indicate differences with the dander detection algorithm. It should be noted that the aim here is not to exhaustively detect the pores or squames present in an area, but rather to evaluate areas of significant density. Indeed, we realize that an exhaustive count requires detection with a very low threshold and therefore the not insignificant appearance of false positives; however, such exhaustive counting is not necessary in the facts.

The general method of pore detection consists of a phase of detection of dark areas, performed on a rate image, and a classification phase, as shown in Figure 11. For dander, the The general method consists of a phase of detection of the light zones, carried out on a rate image and in a classification phase, as shown in FIG.

The main objective of the detection phase is to highlight dark areas on lighter skin (respectively light areas on darker skin for dander). One of the problems here is to correct the problem of non-uniform lighting. Indeed, images may include illuminated areas manifesting as light tasks and may subsequently distort the detection results. To overcome this problem, we use the concept of the image rate (see above). Since a pore is a dark area on a lighter surface, its detection can only be local, that is, locally, in a finite area, a few dark pixels surrounded by lighter pixels are potential pores. We therefore construct the image rate in order to respect differences in contrast locally.

The image rate to solve the problem of the nonuniformity of lighting, the darker points (which correspond to pore peaks) can be detected by a threshold on the image rate, keeping only the points whose value is less than the average of the image, minus a certain number of standard deviations. From the average m of the pixel values of the image and its standard deviation σ, the threshold is taken equal to m - 2.5 * a. Among the pixels of the image whose value is lower than the threshold previously calculated, a thresholding on the size of the possible pores is carried out. Indeed, the sizes greater than a certain number of pixels (81 pixels for example for a resolution of 1664 x 2496 pixels) can not correspond to pores, because they would exceed the normal sizes of a pore, even dilated. Similarly, sizes smaller than 6 pixels are too small to match a pore (for a given size image). Following these two additional thresholds, the areas remaining in the image are dark areas on a lighter surface, and have the dimensions representative of a pore.

The image rate is also used for dander, but thus seeking to detect the clearest points, keeping in the image rate only points whose value is greater than the average value of the image plus 2.5 standard deviations.

The parameters of the algorithm for the phase 1 pore detection are for example the following (values to be adapted according to the shots), for a resolution image of 1664 x 2496 pixels

• Size of the filter 1 to calculate the image rate (5x5)

• Size of the filter 2 to calculate the image rate (9x9)

• Multiplier coefficient of the standard deviation in the thresholding (2,5)

• Maximum size of a pore (Max Size 81 pixels)

• Minimum size of a pore (SizeMin 10 pixels)

And the algorithm of the detection phase can be:

• Read the picture

• turn it into gray level (INDG)

• Calculate the image rate (Itals) from INDG

• Detect pores in Itaux

o Keep only pixel values less than m - 2.5 * σ o Delete guarded areas greater than MaxSize

o Remove guarded areas below SizeMin

After having detected the possible pores, a classification phase according to the shape and the contrast with respect to the neighborhood is carried out. The rule that the overall number of pixels of a pore is less than a certain threshold is sufficient only if the pore has a suitable shape. The pores are generally circular in shape, and in the case where they are dilated, their shape becomes elliptical. According to an internal study, a pore corresponds to an area whose area divided by the square of the perimeter does not exceed 1/6 / r. In the case of an elliptical shape, this threshold implies that the ratio between the major axis and the minor axis is less than 2.61.

Therefore, during this classification step, the area of each previously detected object is modeled as an ellipse. If the ratio of the axes is lower than the threshold above, the zone is kept, otherwise it is rejected. This classification uses color information in objects. In general, the presence of a pore in a zone modifies certain parameters of this zone, depending on whether the pore is present or absent from this same zone. More concretely, a zone containing a pore has an average color value lower than this same zone once the pore has been removed. This same observation remains valid for the median value. Likewise, the darkest value of a pore is supposed to be the darkest value among all the neighborhood around this same pore.

We use advantageously the three preceding rules to decide if a dark area, having the size and shape of a pore is actually a pore. We thus calculate the difference on the mean and the median according to the following diagram:

• Choose a dark area having the shape and size of a pore (after detection phase 1)

• Calculate the median and the average of the pixels on the neighborhood of this zone (itself included)

• Remove this area and calculate the same parameters on the neighborhood alone

• Deduct a relative difference on the median and the average

The parameters of the algorithm for the pore classification phase are for example the following (to be adapted according to the resolution of the image):

• Relative difference in average (-0.005)

• Relative difference on the median (-0.005)

The algorithm of the classification phase can be summarized as follows: for each of the zones kept during phase 1 detection

• Estimate the values of the axes of the ellipse surrounding the area

o If the axis ratio is greater than 2.61, reject the area

• Calculate the differences on the mean and the median, once the pore removed from the zone

o If the deviations are greater than the threshold, reject the zone

• Calculate the darkest pixel in the area

o If it is the darkest in the neighborhood, the area is a pore o Otherwise, reject the area

It should be noted that this classification phase is here applied to each of the R, G, B and NDG (grayscale image) images and, in this case, the final result corresponds to the intersection of the four individual results on each image or only one of these four images (preferably the grayscale image).

The algorithms of the classification phase for dander are analogous. vs. Circles

Dark circles are an extremely widespread cosmetological sign; they are the result of colored variations of the skin, forming regions of color more or less pronounced under the lower eyelids.

There is considerable intra-ethnic variability in the skin color of the volunteers. This makes it useful to develop two distinct ring-ring algorithms, depending on the intensity of the skin color (see above) of the volunteers; two such algorithms are set out below by way of example.

Algorithm "Clear Skin" (or PC)

Once the volunteer has been classified as fair skin, the first step in the ring detection algorithm is a histogram equalization of the G and B bands of the image. Then, the image is transformed from the RGB color space to the space L * a ^* b. Then we calculate the difference of the images a ^* and b, a ^* - b. This result image is then thresholded to detect the lighter pixels (corresponding to the darkest areas of the ring) and the areas conserved after thresholding are cleaned (essentially by morphological transformations such as erosions / dilations). The final area of the ring is then deduced and the final descriptors measured on this area (surface, average intensity in different bands, relative intensity relative to the intensity of the skin without ring, ...).

The various stages are thus for example:

• definitions of areas to be treated (which can be done using a morphological model)

• Histogram equalization on G and B channels (1)

• Transformation into L * a * b

• Calculation of the result image: a ^* - b

• Image Threshold to keep only the lightest pixels

• cleaning of these areas, typically by dilation, erosion and removal of the pupil area and the shadow of the nose, then

• restitution of the result in the form of, for example, an image showing the contours of the ring (see figure 8).

Dark Skin Algorithm (or PF)

As for light skin, we work in the color space L * a * b. On the other hand, the colorimetric segmentation is done on the "b" band of the L ^* a ^* b system. Then, the algorithm is similar to that described above for fair skin. d. stains

It is understood that the notion of spots may differ according to the degree of intensity of skin color (dark spots on light skin, light spots on dark skin, etc.).

In the following, we consider the detection of spots as a dark zone detection on a lighter surface. The detections in the opposite case will obey the same detection / classification scheme with suitable parameters. We will not present this case in detail.

A possible algorithm for spot detection consists of two phases, one phase of dark area detection, and another of classification of these areas into spots. The next paragraph is devoted to the presentation of this method. An example of an overall scheme of the method for dark skin and for fair skin is presented in Figure 13 and in the rest of this text.

From an initial image with spots, we calculate the rate image. On it, the outline of the spot is darker than the rest of the spot itself, and therefore, the rate image only detects fragments of the spot instead of the complete spot. We need to use an algorithm to link the stain areas to fully detect them. An algorithm called "active contours" can solve this kind of problems.

An active contour is a contour that automatically adapts to a structure based on a given criterion, such as the color of the image (in our case). From an initial contour, it is necessary to modify the latter in an iterative manner so that the final contour represents a homogeneous zone, of given color, for example.

Several active contours methods exist, among which we opt for the "Level Set" methods. It involves minimizing an "energy" function, the sum of "internal energy" (which respects certain mathematical criteria) and "external energy" (represented by the length and area of the area). The minimization of this energy makes the algorithm converge towards the optimal zone.

We apply this method to the contours detected on the rate image. In order to improve the performance of the detection, a contrast enhancement is necessary in the case of fair skin. For dark-skinned volunteers, we do without this step, as the performance of the algorithm is better without contrast enhancement.

The algorithm of the detection phase is therefore the following:

Parameters (values to be adapted according to the shots)

• Size of the filters to calculate the image rate

• Threshold on dark values of the image rate (10%)

• Active Contour Algorithm Parameters Algorithm

• Read the picture

• If clear skin, enhance the contrast

• Choose the B component of the RGB image (enhanced if clear skin)

• Calculate the rate image from image B

• Detect stains in Image rate

o Keep only the 10% of the darkest pixels in the image rate

• Launch the active contours algorithm

• The remaining areas are potentially stains (to be validated in phase 2)

After detecting the possible stains, a classification phase is carried out, and is based on the comparison of the contrast with respect to the neighborhood, and on the criteria of shape and presence of pores. This step is similar to that used for the detection of pores and dander, and is based on the color differences between the spots and their neighborhood. This assumes that a spot is darker than its neighborhood, and that it is sufficient to calculate the following two terms:

• the first term is a difference between the statistical distributions of the task and its surroundings. Indeed, a spot being darker, its distribution is shifted compared to that of its neighborhood, which implies a distribution gap more consequent when a stain is well marked.

• the second term is a ratio between the average color of the area including the spot and that of its neighborhood (having removed the stain). The spot being darker than its neighborhood, this term is less than 1 if it is indeed a spot (darker than its neighborhood).

The algorithm of classification according to the contrast is thus the following one:

Settings (to be adapted according to the resolution of the image and the shooting conditions)

• Report of medium colors (0.99) • the threshold on the difference between the distributions is calculated automatically, according to the zones rejected by the report of the colors

Algorithm; For each of the previously selected tasks:

• calculate the difference between the two distributions (of the spot and its neighborhood), and the ratio of averages

• Remove spots with a ratio of the average above the threshold

• For these deleted tasks, calculate the average of the differences between the distributions, and choose this average value as threshold

• Remove stains with a difference between distributions above this threshold

Finally there is a classification according to the form and the presence of the spots. This step assumes that a spot has an elliptical shape and assumes a threshold similar to that used for the detection of pores and dander (eg 3.06, threshold obtained experimentally). In some cases, especially at the forehead, the existence of pores can bias the task validation algorithm. Indeed, a succession of pores can induce zones for which the criteria of color and form are similar to those observed for the real spots (the distribution of a zone is global and not local, the average value is also ). In order to solve this problem, we use the pore detection algorithm to identify the number of pores and the relative area they occupy in each zone. At the end of this step, the remaining areas correspond to the spots to which the algorithm has converged.

The algorithm of morphological classification is thus the following one: Parameters (to be adapted according to the resolution of the image and the conditions of shooting)

• Maximum size of the spot, in pixels

• Ratio of the axes of the ellipse circumscribed to the spot (3.06)

• The maximum number of pores that a spot can include (2 pores) • The maximum number of pores beyond which the stain is rejected (5 pores)

• The maximum relative area of pores in a stain (15%)

Algorithm: for each of the previously selected tasks:

• calculate the ratio of the axes of the circumscribed ellipse, and delete the task if the ratio is greater than the threshold

• Calculate the area of the study area, and delete it if it exceeds the threshold in question

• Count the number of pores in the spot

o If this number is less than or equal to 2, keep the task

o If this number is greater than or equal to 6, delete the task

• Otherwise, calculate the relative area, and keep the spot only if the area is below the threshold in question.

• It is sometimes necessary to apply a mask on the image such as to remove the area of the nose, or eyebrows, etc. e. Télanqiectasies

Telangiectasia refers to an abnormal vascular dilatation in size and permanence. It is red, non-pulsatile and forms a fine line, tortuous, often in arborization or network and is most often located, especially in the face. It is understood that this type of cutaneous sign may not exist or not be detectable on dark skin.

From the color image (components R, G and B), a contrast enhancement is performed. This enhanced image makes it easier to visualize the vessels and is the initial image.

As an example, at first, we transform this image of the RGB space into the space L ^* a ^* b, where we apply a median filter to smooth the image. The purplish-red colors are especially visible in the channel "a". From this image that we normalize, we apply a filter "top-hat", to highlight the light areas on a dark background. The image is thresholded to obtain a first coarse detection image of telangiectasia. In a second step, we make a color segmentation from the RGB image obtained after contrast enhancement. Only the region where the average value in "a" is maximum is then kept.

We then obtain a second image of coarse detection of telangiectasia.

We then combine these two images. The regions that are not telangiectasias are then cleaned. Indeed, wrinkles and spots are also detected by this algorithm, detections that must not be taken into account. For this, we eliminate the rounded regions, which are usually spots. A mask is also applied not to take into account the regions located at the level of the nasolabial fold. Finally, to keep only the telangiectasias, we look for each region detected, the R, G and B values of each pixel. If the pixel is rather brown, it is removed. If it is rather purple-red, keep it. To determine this threshold that eliminates the pixels near the spots (dark or light) at the level of colorimetry, we have collected a database containing the R, G and B values for a certain number of pixels, from images obtained after enhancement of contrast. On the one hand, we have R, G and B values corresponding to spots, and on the other R, G and B values corresponding to telangiectasia. When doing a Principal Component Analysis (PCA) on this database, we clearly see a separation between spots and telangiectasias (see Figure 14). f. Rides

The types of wrinkles involved include the following:

• forehead expression lines

• lion's wrinkles

• wrinkles of the crow's feet

• wrinkles under the eyes

• the nasolabial fold

• lip contour lines

• the wrinkles of the corner of the mouth Detection Algorithm

By way of example, to detect these signs, the algorithms used are mainly based on "curvelets" (see: Candès E, and Donoho D., Curvelets A surprisingly effective nonadaptative representation for objects with edges, curves and surfaces, Curves and Surfaces 1999). These are a specialized version of the family of anisotropic wavelets that are very well adapted to the representation of discontinuities along the contours.

The parameters used for curvelets, as well as the thresholds for the extraction of wrinkles are a function of the size (coarse or fine) of the wrinkle to be detected. A post-treatment is then generally useful according to the orientation of the wrinkle, to isolate it from less interesting structures.

All the algorithms for detecting wrinkles can follow the same diagram explained below:

• calculate the transform into curvelets on the image to be treated I,

• Denoise the image I by thresholding the images in the space of the curvelets, with a parameter appropriate to each type of ride (sigma noise standard deviation). We get the image an image "the",

• apply a morphological treatment (top hat or bottom hat according to the image used initially) to highlight the contours of interest. We obtain the image "lem",

• threshold the image lem, to produce a binary image containing only the structures previously highlighted. The value of the threshold varies according to the type of wrinkles to be detected, and is of the average type (lcm) + k * standard deviation (lcm). Let "lems" be the threshold image,

• then apply a post-treatment to eliminate the elements of the image that do not correspond to the type of wrinkles sought. The selection criteria include, among others, size, orientation and shape. The remaining structures are considered as wrinkles and their descriptors are calculated. Nature Descriptors depend on the type of wrinkles but can range from size, length or average intensity.

The general algorithm summarized above is to adapt to each type of ride. The adaptation parameters are as follows:

The image I on which the curvelet transform and the following operations are performed (for example: image L, image b, etc.)

• The morphological operation performed before thresholding on the transformed image I

• The size of the structuring element of this morphological operation

• sigma (denoising parameter in curvelet)

• k (parameter giving the value of the threshold for final thresholding).

Classification algorithms

In general, the choice of classification method depends on the data available. In our case, since we have images and associated grades, it is therefore more interesting to use supervised classification methods. The classification methodology presented here as an example is based on two pillars:

1. Extraction of image descriptors

The descriptors, also called characteristics or parameters, are a set of measurements obtained from the image, which make it possible to describe it or to characterize it. They describe the content of the image and thus make it possible to identify it.

2. Supervised classifications

The latter have the strategy of using a set of samples (called learning) to learn the classification parameters (and build a model), and test them on another set (called test) to define the quality of the classification. This is the prediction phase, which consists of using the built model to assign a class to a new image.

As an example, we show below the application of the SVM (support vector machinery) classification method, which is based on on a criterion of maximizing the separation margin between classes, on the descriptors drawn from images of lion's wrinkles. The type of descriptors used is important because the success of subsequent operations depends on this information extracted from the image.

Identification and extraction of descriptors in the case of lion wrinkles

The criteria we need to classify lion wrinkles, for example, are not necessarily the same as those we will use for stains. In addition, for each criterion, the descriptors must be relevant enough to recognize all images of the same grade, and at the same time discriminating enough so that there is no confusion with other classes.

In practice, on the wrinkles of the lion it is found that according to the grades, the number of wrinkles varies, as well as the length and thickness of these wrinkles. We can therefore deduce that these characteristics will have to be taken into account in order to hope to be able to separate the different grades of the lion's wrinkles.

In the present study, we have mainly used spectral descriptors (gray levels of different spectral bands, histograms) and geometric descriptors (shape measurements). A summary of extracted descriptors for lion's wrinkles is given below:

- number of wrinkles,

- thicknesses of the two longest wrinkles,

- lengths of the two longest wrinkles.

Selection of descriptors

After the extraction procedure of the descriptors, the image is thus represented as a set of these characteristics. These are often combined or concatenated to improve the performance of the classification. It is important to take into account the dimensionality of the data, ie the number of characteristics, the latter being able to influence the classification results significantly. Indeed, the performance of the classifiers does not increase indefinitely with the size of the vector descriptors. Moreover, the complexity of the classification, in terms of computation time, increases with the size of the feature vector. It is therefore interesting to limit the number of effective descriptors to the "optimal" number, by selecting the most relevant descriptors. This selection of descriptors can be operated in several ways, depending on the classification method used. For example, selecting a relevant subset from a set of descriptors is done by calculating a score for each descriptor, based on the values of these descriptors, and the number of positive and negative examples. This results in a descriptor scheduling, which keeps only the first n descriptors that give the best performance when evaluating the model generated during the learning phase.

Image taking wave lengths

For most signs, we can associate a spectral band in which this sign has a maximum signal-to-noise ratio.

The wavelengths indicated in the following table are those in which a criterion is the most visible (we can see this criterion outside these bands but its contrast is much lower).

It should be noted that the wavelength of 570 nanometers could also be suitable for darker skin, instead of the 620 nanometer frequency (it would be just below the min threshold for dark circles, but would be well located in the other defects); it follows that, being limited to barely two wavelengths in the order of 500 nanometers and 570 nanometers, it is possible to detect the types of cutaneous signs mentioned above. However, alternatively, the frequencies of 620 nanometers and 570 nanometers can be combined for the darker skin and to treat those of the images which allow an optimal contrast; Thus, advantageously, there are at least three wavelength ranges, here 500 nm (+/- 20 nm), 570 nm (+/- 20 nm) and 620 nm (+/- - 20 nm), for the detection of cutaneous signs whatever the degree of intensity of skin color.

The spots (of which lentigines and ephelids are examples) are traced mainly by the absorption of melanin (dominant decreasing exponential around 500nm for light skin). On the other hand, for dark skin, the massive presence of melanin throughout the skin as well as the degradation of the signal on noise towards 500nm displaces the bands of visibility of the spots towards 560-630nm.

Dark circles are very visible around the peaks of hemoglobin. This can probably be explained by the thinness of the skin at the rings, and therefore better absorption of blood vessels.

It is recalled that a chromophore is a molecule that absorbs electromagnetic energy at a given wavelength, with a characteristic yield given by the extinction coefficient of the molecule. For example, carotene is the chromophore that gives color to many fruits (such as carrots), this molecule actually absorbs wavelengths in the blue range of the visible spectrum and therefore only returns the complementary color (orange and white). red). Melanin and hemoglobins are the most important chromophores for understanding how the skin works.

Melanin in the skin is the dominant chromophore of the epidermis. There are two types of melanin pigments: Peumelanin and Pheomelanin.

Hemoglobin is a red chromophore found mainly in red blood cells. When hemoglobin contains oxygen, it is called oxyhemoglobin. In the opposite case, it is called deoxyhemoglobin. As a first approximation, the epidermis can be seen as a layer of melanin and the dermis as a layer of hemoglobin. Therefore, the skin color will depend on the variation of the amounts of hemoglobin and melanin.

Oxyhemoglobin and deoxyhemoglobin have characteristic absorption peaks, and absorption maxima for oxyhemoglobin and deoxyhemoglobin are observed around 430 nm and 550 nm in the visible spectrum.

By way of example, an implementation system of the invention (such as that of FIG. 2 or 3) comprises:

• An image capture device with

A sensor of 5 Mpixels

A focal length (for example fixed and without autofocus)

• regularly distributed LEDs, having the following wavelengths

·. λ ~ 500 (+/- 20) nm, that is to say a blue-green hue,

·. λ -570 (+/- 20) nm, that is to say an orange-yellow hue, ·. λ -620 (+/- 20) nm, i.e. a red hue,

• white LEDs regularly distributed.

The size of the sensor can be increased to 10 Mpixels without having unacceptable processing times. Zoom effects allow conversely to go down to resolutions of 2 Mpixels.

It will be noted that, in the foregoing, several original points per se can be identified as follows:

• Automatic pre-classification of faces according to two criteria:

The algorithms are adapted to different skin colors (either via different algorithms, or via the same algorithms that take different images as input, or via algorithms that have different parameters for different skin colors). Thanks to the invention, a method is available which makes it possible to acquire an image of a new subject and to automatically select the algorithm without any intervention of an operator: it is automatically classified by the results of the 'ACP for the degree of intensity of the color of skin and therefore the choice of algorithm or parameters to use is also automatic.

The criteria used for the pre-classification are:

o Level of Intensity of Skin Color (Mainly): A study using PCA (Principal Component Analysis) was performed on cheek areas of 470 images from the photographic database and identified a threshold between the light skins of dark skin,

o Morphotype (alternatively): an indexing (ie measuring the characteristic points of a face, the indexing can be manual, or preferably automatic) of the majority of the images of the database Has been done. The information contained in this indexing makes it possible to classify the faces according to their morphotype and thus to be able to easily clear the characteristic zones of the face (forehead, nose, eyes, mouth, chin, ears, ...), which can contribute to facilitate automatic processing.

The fact of using mainly the first of these criteria, possibly in combination with the second, to pre-classify faces and automatically process them with different algorithms brings quite significant advantages.

"Rate" image: the "rate" image is introduced to detect pores, scales, spots and telangiectasias. This image is the result of the ratio between two filtered images: in each pixel, it replaces the value of this pixel by the ratio between the average of the pixels of an area directly surrounding the pixel and the average of the pixels of the neighborhood of this zone. If the size of the area directly surrounding a given pixel is characteristic of a structure to be detected, the contrast of this structure will be enhanced. In addition, this transformation being local, it makes it possible to overcome the non-uniformity of the lighting of an image. Curvelets for the detection of wrinkles: use of curvelets to detect wrinkles: generalization of wavelets, they use basic functions with a spatial location, a scale (like wavelets) but also a direction. This gives them a great sensitivity to detect long and thin structures and allow to choose only the structures having a given direction (interesting for forehead lines that are horizontal, for example or for most wrinkles that have a given direction ).

Use of SVM for grade classification: Use of SVM (vector support machine) type methods for classification into grades after extraction of descriptors. These methods, which require learning, make it possible to classify in a probabilistic way the descriptors derived from the detection. It does not seem to us that they have been used yet to discriminate cutaneous grades.

Application of a differential image acquisition strategy: to avoid problems caused by an uncontrolled luminous environment, this strategy consists in taking two consecutive images, one illuminated with ambient lighting and other with ambient lighting and controlled lighting. Then we subtract the second from the first and we obtain an image in a controlled light environment.

More generally, the invention also proposes an association of hardware and software to perform a task for cosmetology, by implementing an apparatus that acquires images in a differential manner by illuminating, at the desired time, the face to be treated and which, Automatically, calculates grades for certain cutaneous signs using algorithms whose parameters are automatically selected according to the type of skin of the face, which brings significant advantages.

Claims

1. A method for detecting and quantifying cutaneous signs on a skin zone, according to which:

^* Is identified in advance for a plurality of at least two classes of degrees of predetermined skin color intensity and a plurality of predetermined types of skin lesions, a plurality of wavelengths selected so that, in each combining any type of skin sign within said plurality of types and any class of skin color intensity level within said plurality of classes, at least one of the wavelengths of the plurality of wavelengths allows the contrast detection of such a type of sign on the skin of a person having a degree of skin color intensity within said class,

^* we take under white lighting at least one image of a selected area of the face of a subject,

is automatically identified from this image the degree of intensity of skin color corresponding to the person concerned and identifies the class to which this degree belongs,

the skin or wavelengths which, for this degree of intensity of skin color, allow the detection of the skin to be detected, is automatically selected according to this class of degree of skin color intensity; a type of sign that one seeks to detect on the chosen zone of the face,

^* we take at least one monochromatic image at the selected wavelength, which is treated with an algorithm and parameters mainly chosen according to the class of degree of intensity of skin color, in sort of detecting and grading signs of said type that one seeks to detect.

2. Method according to claim 1, characterized in that the plurality of wavelengths comprises wavelengths of the order of 500 nm (+/- 20 nm) and 570nm (+/- 20 nm).

3. Method according to claim 1 or claim 2, characterized in that this plurality of wavelengths is at most equal to three for all classes of degrees of intensity of skin color.

4. The method of claim 2 or claim 3, characterized in that the plurality of wavelengths further comprises a wavelength of the order of 620 nm (+/- 20 nm).

5. Method according to any one of claims 1 to 4, characterized in that the plurality of types of cutaneous signs comprises wrinkles, dark circles and pigment irregularities.

6. Method according to any one of claims 1 to 5, characterized in that, for at least one of the classes of degrees of intensity of skin, one chooses two wavelengths for the same type of cutaneous sign.

7. Method according to any one of claims 1 to 6, characterized in that the plurality of wavelengths is composed of three wavelengths and the plurality of classes of intensity levels of skin color consists of two degrees, namely light skin and dark skin.

8. Method according to claim 7, characterized in that one affects at most two wavelengths for the types of cutaneous signs for light skin, and at most two wavelengths for the types of cutaneous signs for the skin. dark skin.

9. Method according to any one of claims 1 to 8, characterized in that the images at a given wavelength are taken by difference between a first raw image taken with a given ambient lighting plus a lighting specific to this given wavelength and a second raw image of the same skin zone taken with this given ambient lighting only.

10. Method according to any one of claims 1 to 9, characterized in that one captures a plurality of images for the same wavelength, at increasing levels of lighting power, and selects the image obtained with the highest power level without saturation.

11. A device for detecting and quantifying cutaneous signs on areas of skin, comprising: an illumination device adapted to illuminate a skin zone placed at a given location under several illumination regimes, as desired,

^* a treatment device,

this treatment device:

containing identification data, for a plurality of at least two classes of predetermined skin color intensity levels and a plurality of predetermined skin sign types, of a plurality of selected wavelengths so that, for each combination of any type of skin sign within the plurality of types and any class of skin color intensity degree within the plurality of classes, one of the wavelengths of the plurality of wavelengths allows the contrast detection of such a type of sign on the skin of a person having a degree of skin color intensity within said class,

being designed so as to cooperate with the capture device and the illumination device for capturing images of the skin zone chosen in white light and at said wavelengths of said plurality, the illumination regimes permitting such capturing in white light and capturing at each of the wavelengths of this plurality, and

being designed so as to automatically identify in at least one image captured by the white-light image-capturing device the degree of skin-color intensity corresponding to the person concerned and the class to which that degree belongs of intensity of skin color, to select, automatically according to this class of degree of intensity of skin color, that of the wavelengths which allows, for this degree of intensity of skin color, the detection of a type of sign that one seeks to detect, to automatically capture at least one image at least this selected wavelength and to automatically process at least this captured image at this selected wavelength using an algorithm and parameters mainly chosen according to the class of degree of intensity of skin color, so as to detect signs of the type that one seeks to detect.

Device according to claim 11, characterized in that the illumination device comprises a plurality of light sources arranged around the image-capturing device and having one or the other of the chosen wavelengths or being a white light source.

13. Device according to claim 12, characterized in that the plurality of light sources is disposed on a concave curved surface whose concavity is turned towards said location.

14. Device according to any one of claims 11 to 13, characterized in that the illumination device comprises a plurality of elementary light sources having wavelengths of the order of 500 nm (+/- 20 nm) , 570 nm (+/- 20 nm) and 620 nm (+/- 20 nm).