LIP MODELING FROM 2D AND 3D IMAGES

Abstract

LIP MODELING FROM 2D AND 3D IMAGES The invention relates to a method for manufacturing a personalized applicator for applying a cosmetic composition to the lips. This applicator comprises an application surface made of a material capable of absorbing the composition. The method comprises the following steps: Capturing a 2D input image with a first dimensional reference frame; Capturing a 3D input image with a second dimensional reference frame. Figure 10

Description

Title of the invention: LIP MODELING FROM 2D AND 3D IMAGES

[0001] The present invention relates to methods for manufacturing a personalized applicator for applying a cosmetic composition to the lips. The invention further relates to the personalized applicators thus manufactured and to cosmetic treatment methods, particularly makeup application, using them.

[0002] More generally, a cosmetic product is a product as defined in Regulation (EC) No 1223 / 2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products. Technological background

[0003] The usual method for applying lip makeup involves applying a film of covering and / or tinted product using an applicator such as a lip brush or lipstick, which is moved along the lips to cover their surface. The user observes the effect but may be dissatisfied with the result. In particular, if they feel that the shape of their lips is not to their liking, the user remains disappointed with the result. This is the case, for example, for people who consider their lips to be too thin, too wide, asymmetrical, or poorly proportioned in relation to the shape of their face.

[0004] Users generally desire a clean lip look and prefer to use lipstick. Unfortunately, lipstick is poorly suited to creating precise, error-free contours, and using a lip pencil is not always easy, especially when one does not want to follow the natural lip line.

[0005] It is known from documents FR 752 860, US 2 279 781, US 2 207 959, FR 663 805, US 2 412 073, US 3 308 837, US 2 735 435, US 1 556 744, US 2 248 533, US 2 416 029, US 1 944 691, US 1 782 911, US 2 554 965, US 2 199 720, WO 2008 / 013608, US 2003 / 209254 and US 2010 / 0322693 to produce an applicator whose application surface has the predetermined shape of a mouth. This solution allows for standard makeup application but is not very satisfactory as it does not always conform to the three-dimensional morphology of the lips, leaving areas uncovered.

[0006] Other solutions have been described in applications WO 2013 / 045332, FR 2 980 345, WO 2013 / 092726, and FR 2 984 699 to create an applicator adapted to the individual morphology of the lips. To do this, an impression of the user's lips is taken from a generically corrected lip contour scan, then a counter-mold is made, which will serve as the applicator. The user places the product in the counter-mold before applying it to their lips. Another One option involves delivering the product through a counter-mold via numerous holes. This solution offers improvements, particularly in terms of precision and speed of application, but it doesn't allow for ideal makeup application, and the applicator doesn't conform sufficiently to the three-dimensional shape of the lips.

[0007] These solutions also do not allow for very rapid availability of the applicator for the user.

[0008] To overcome these drawbacks, patent application EP3697257A1, filed by the Applicant, describes a method for manufacturing a customized applicator for applying a cosmetic composition to the lips. This method involves performing an input 3D scan of the topography of a portion of the lip surface and manufacturing the applicator by machining a preform or by additive manufacturing, based on the scan. The document also discloses the estimation of the natural lip contour from an image of the lips, without specifying whether this is a 2D input image or a complementary 3D input image. In any event, it does not describe the implementation of a first 2D input image with a first dimensional reference frame, nor the implementation of a second 2D input image with a second dimensional reference frame.

[0009] The article “3D Shape Estimation from 2D Landmarks: A Convex Relaxation Approach”, in Proceedings of CVPR 2015, by Xiaowei Zhou, Spyridon Leonardos, Xiaoyan Hu, and Kostas Daniilidis, Computer Vision and Pattern Recognition, investigates the problem of estimating the 3D shape of an object, given a set of 2D landmarks in a single image. The developed method is applied to estimating the shape of a human pose (full body of a standing person) and the shape of a car. The disclosed method relies on examining an input 2D photograph of the person or car, equipped with a reference frame. An output 3D shape is estimated from a convex formulation derived from the input 2D photograph and an algorithm for solving the convex program. However, this article does not provide for 3D modeling of the lips, nor for the creation of a lip applicator.Furthermore, it discloses a relatively complex algorithm dedicated to the views studied, based on a convex program. Finally, it does not disclose the study of reference points from the 3D input image.

[0010] In the specific case of lips, a major problem lies in the fact that:

[0011] The 2D image of lips is necessarily imperfect, since it depends on the lighting during photography, the viewing angle, makeup, and the surface condition of the lips, particularly their roughness or texture. Consequently, it is not possible to extract reliable dimensions from a 2D input image alone that could be used to create a personalized applicator perfectly adapted to a user's lips.

[0012] The same applies to an input 3D image of the lips, the result of which depends on the skin's surface texture, the lighting, the person's exact positioning, their facial expressions or gestures, and even the imperceptible movement of facial muscles, including those of the lips, during the image capture. Consequently, dimensions extrapolated from a 3D scan will inevitably lack accuracy, and it follows that truly reliable dimensions cannot be extracted from this 3D image alone to create a personalized applicator perfectly adapted to a user's lips.

[0013] The problem posed by the invention is to propose a method for modeling the lips and manufacturing a personalized makeup applicator that is a reliable method, that is to say, providing a model of the lips that is more conforming and precise than existing models, in order to produce a perfectly personalized applicator, and this despite the context specifically related to the lips, animated parts of a person, including imperceptible movements or expressions, sometimes involuntary.

[0014] Furthermore, the method must be simple, both in terms of its input parameters and the algorithm implemented for analyzing and processing those input parameters. Definition of the invention

[0015] The invention relates to a method for manufacturing a personalized applicator for applying a cosmetic composition to the lips, this applicator comprising an application surface made of a material capable of absorbing the composition, characterized in that the method comprises the following steps:

[0016] To capture a 2D input image equipped with a first dimensional reference frame, of at least a part of the surface of the lips,

[0017] To capture a 3D input image equipped with a second dimensional reference frame, of at least part of the lip surface, and

[0018] To produce, from at least the input 2D image, equipped with the first dimensional reference frame, and the input 3D image, equipped with the second dimensional reference frame, at least a part of the applicator or of a mold used for its manufacture, by machining a preform or by additive manufacturing.

[0019] The method according to the invention uses a 2D input image equipped with a dimensional coordinate system and a 3D input image equipped with a dimensional coordinate system. Thanks to these two dimensional coordinate systems, it becomes possible to improve the dimensional accuracy of the applicator or the mold used to manufacture it, and to get as close as possible to the actual dimensions of the person's lips.

[0020] The applicator resulting from the process according to the invention is more faithful and conforms more closely to the lips than those of the prior art. As a result, the personalization of the applicator is more successful, for better makeup application.

[0021] The invention allows for the self-application of professional makeup on a surface of the lips, using a morphological applicator adapted to the user and made to measure.

[0022] The applicator customized according to the invention makes it possible in particular to draw by to fill the mouth, and to color it evenly if desired.

[0023] The invention also relates to a method of lip makeup, comprising the application of a cosmetic composition to the lips using an applicator obtained by the method described above.

[0024] The invention makes it possible to achieve a makeup result with a clean contour, enhancing the harmony of the face. The invention also offers a way to apply makeup very quickly, in a single step, and anywhere, even without a mirror, for example in the car or at the office.

[0025] The invention makes it possible to apply makeup with bright colors and / or with a high-wearing composition without risk, even on a daily basis, because the personalized applicator makes it possible to avoid the mistakes caused by this type of product when applied with known applicators.

[0026] The customized applicator according to the invention makes it possible to redraw the contour of the lips, by providing a remodeling effect, and can therefore be used by people whose contour has become blurred, particularly due to skin aging, and who no longer dared to wear makeup.

[0027] The invention also relates to a computer modeling method for at least one area of ​​the lips, the method comprising the operations of:

[0028] To capture a 2D input image equipped with a first dimensional reference frame, of at least a part of the surface of the lips,

[0029] To capture a 3D input image equipped with a second dimensional reference frame, of at least part of the lip surface, and

[0030] Generate a 3D output image of the lip area from the input 2D image and the input 3D image, the lip contour.

[0031] Another object of the invention relates to a computer 3D modeling system for at least one area of ​​the lips, preferably intended for use in the manufacture of a personalized applicator for applying a cosmetic product to the lips as defined above, the system comprising at least one mobile 2D and 3D image capture device, in particular a smartphone, in which, once the mobile image capture device is placed in a predetermined position relative to the lips, the mobile image capture device is capable of capturing a 2D input image, equipped with a first dimensional reference frame, of at least a part of the lip surface, and of capturing a 3D input image, equipped with a second dimensional reference frame, of at least the same part of the lip surface. including further a processor capable of generating a 3D output image of the lip area from the 2D input image and the 3D input image, by determining the lip contour of the 3D output image from the 2D input image.

[0032] Another object of the invention relates to a customized applicator for applying a cosmetic composition to the lips, obtainable by the process described above. This applicator differs from existing applicators in its degree of precision in conforming to the shape of the lips, conforming more closely to the surface of the lips with greater accuracy.

[0033] Input 2D image

[0034] Preferably, a smartphone is used as the input 2D image capture device.

[0035] Preferably, said mobile image capture device is placed on the same plane as the area to be measured, in particular vertically and under the user's chin to measure an area of ​​his face, in particular to record the dimensions of his lips.

[0036] Information on the positioning of the mobile image capture device relative to the area to be measured can be provided, in particular by using a position sensor, in particular an accelerometer or a gyroscope.

[0037] The image is preferably captured automatically by the mobile image capture device once the said predetermined position relative to the area to be measured is reached.

[0038] One or more signals, including voice signals, can be sent to the user to help them position the mobile image capture device in the predetermined position relative to the area to be measured and / or relative to the mirror. Electronic guidance of the user is thus possible.

[0039] Reworked surface

[0040] By "reworked surface" we mean a surface whose shape and / or contour has been modified in relation to the natural surface whose topography has been acquired.

[0041] The method according to the invention may thus include a step consisting of generating a reworked 3D surface from the data from the acquisition of the surface topography, in particular using image processing software.

[0042] In particular, a reworked 3D surface can be generated by stretching the lip contour obtained from the input 2D image. The input 2D image can be reworked.

[0043] The reworked surface may optionally deviate from the natural surface of the lips within their contour to create a space between the application surface and the scanned lips, when the applicator is applied normally to The lips. This space can be used to accommodate a self-expanding composition, as will be detailed later.

[0044] The reworked surface can coincide with the natural surface of the lips resulting from the scan, except for its contour which is different from the natural contour of the scanned lips, to modify the contour of the made-up lips.

[0045] The method according to the invention may include a step consisting of giving a user the possibility of choosing between at least two makeup results, the reworked surface being generated at least from this choice, for example automatically by means of software.

[0046] The method according to the invention may include the step of allowing a user to model a surface obtained from the input 2D image, in particular its contour, and thus generate the reworked surface. The modeling can be performed remotely via software from a workstation to which data representing the 3D image has been transmitted via a telecommunications network, in particular the Internet or GSM / GPRS network. This remote workstation is, for example, that of a makeup artist.

[0047] Input 3D image

[0048] To create an input 3D image, any 3D scanner capable of capturing the volume and dimensions of the area in question can be used. Preferably, a 3D scanner capable of also capturing the color and appearance of the area in question is used, so as to acquire one or more images providing information on the location of the composition.

[0049] The input 3D scan is advantageously a light fringe projection scan, but any other structured light is possible.

[0050] The input 3D image can be acquired from the camera of a mobile phone (smartphone).

[0051] However, the invention is not limited to a particular type of mobile image capture device.

[0052] The mobile image capture device, particularly in the case of a smartphone, may include one or more microprocessors or microcontrollers and additional circuits, arranged to execute an application aimed at carrying out the steps of the process according to the invention.

[0053] The detected reference points and / or measured dimensions are advantageously stored on the mobile image capture device and / or transmitted to a remote workstation connected to the mobile image capture device for the manufacture of a customized applicator. This enables rapid and reliable manufacturing.

[0054] A specific pattern can be displayed on the screen of the mobile image capture device, in the form of specific shapes or displays similar to QR codes, During image acquisition, to increase the robustness of screen detection, an error message may appear to instruct the user to re-acquire the image under better conditions, particularly if there is a light reflection on the screen of the mobile image capture, partially obscuring the specific pattern. This is especially important if there is a light reflection on the screen of the mobile image capture, which may partially obscure the specific pattern.

[0055] In the case of the lips, the possible measured dimensions may be their length between the two extreme points, the height of the upper lip at different points along the lips, the height of the Cupid's bow, the distance between the philtrum crest and the midline of the lip, the height of the lips at different points along the lips, the distance from the corner of the lips to different points along the lips and / or the height of the commissure from the highest point of the lips.

[0056] According to the invention, the input 2D image and / or the input 3D image and / or the output 3D image can be reworked.

[0057] Thus, the method according to the invention may include establishing a remote link with a third party providing a model to be offered to the person whose lips have been scanned according to the physiognomy of that person, for example via an Internet video conferencing platform.

[0058] The method according to the invention may include the detection, in particular automatically by means of software, of an asymmetry of the lips and / or face; the calculation of the reworked surface may be carried out, preferably automatically, at least as a function of the detected asymmetry.

[0059] Manufacture of the applicator

[0060] A file readable by a CNC machining center or a 3D printer is advantageously generated, which can be stored, in particular automatically, for example in the cloud or on a central server, and can be sent to all user access points, for example points of sale or institutes. The file can be transmitted to the user. It is also possible to retain the rejected files, in order to avoid redundant testing.

[0061] A translated digital copy of a surface, possibly reworked, obtained from the 3D scan of the lips is advantageously created, and then a smoothed volume of the applicator or mold between said surface and its translated copy can be generated. In an alternative embodiment, a smoothed volume of the applicator or mold is generated between said surface and a standard applicator surface, in particular created by learning from several acquired surfaces.

[0062] The applicator can be produced by machining, preferably by micromachining. Advantageously, a preform is machined from a selection, particularly automated, chosen from several options depending on the shape to be obtained after machining. This reduces manufacturing time. These preforms may have been custom-made, for example, from of different mouths, and their machining face is advantageously larger than the surface of natural lips. The preforms may have the reverse side already formed, with or without a handle, or with or without a system for adding a handle, or with or without a system for adding a compartment that can hold a cosmetic product.

[0063] The invention offers, if desired, the possibility of remotely reproducing the applicator, either when one is traveling and has forgotten it, or because it has been lost, or because one wishes to share the applicator with another person. It is sufficient to send or have sent the 3D file, stored in computer memory, to create a reproduction anywhere.

[0064] The 3D printer can be a so-called filament printer. The 3D printer used can achieve a z-axis accuracy of 0.5mm, better 0.1mm, even better 0.03mm.

[0065] In the case of 3D printing, printing can be carried out on a support or on a predetermined object, such as, for example, a preform with or without a handle, with or without a system for attaching a handle, or with or without a compartment that can contain a cosmetic product. Preferred modes of implementation

[0066] Preferably, the application element according to the invention has one or more of the following characteristics, taken alone or in combination:

[0067] Method for manufacturing a custom applicator

[0068] It includes the determination of a plurality of points of the lip contour, from the input 2D image, and the estimation of the lip contour on the 3D output image, by interpolation from these points.

[0069] It includes determining the depth of the lips on the output 3D image from the input 3D image.

[0070] It includes the detection, on the input 2D image, of several first reference points delimiting the contour of the lips and of several second reference points located on either side of the line separating the lips, in order to produce the contour of the lips on the output 3D image.

[0071] It includes the detection, on the input 3D image, of several third reference points delimiting the corners of the lips and of several fourth reference points located on the longitudinal axis X of the lips, in order to produce the depth of the lips on the output 3D image.

[0072] It includes determining at least one reference point visible both on the input 2D image and on the input 3D image and assigning this reference point a dimensional coordinate on an output 3D image.

[0073] It includes the display of a printable and / or manufacturable output 3D image.

[0074] It includes positioning information for a mobile capture device images in relation to the lip area, notably by using a position sensor.

[0075] It includes the generation of a reworked 3D output surface, in particular by stretching the 2D input image, the applicator or mold used for its manufacture having a shape given at least in part by this reworked surface.

[0076] It includes determining at least one reference point visible both on the input 2D image and on the input 3D image and assigning to this reference point a dimensional coordinate of the output 3D image.

[0077] Lip makeup process

[0078] It includes determining the lip contour of the output 3D image from the input 2D image.

[0079] It includes determining the lip depth of the output 3D image from the input 3D image.

[0080] It includes positioning information of a mobile image capture device relative to the lip area, in particular using a position sensor, in particular an accelerometer or a gyroscope.

[0081] It includes the generation of a reworked 3D output surface, in particular by stretching the 2D input image, the applicator or mold used for its manufacture having a shape given at least in part by this reworked surface. Description of the figures

[0082] Other features and advantages of the invention will become apparent from the following detailed description, non-limiting examples of its implementation, and from examination of the accompanying drawing, in which:

[0083] [Fig.1]

[0084] Figure 1 represents a set of reference points detected on an input 2D image by analyzing the image with 2D image analysis software,

[0085] [Fig.2]

[0086] Figure 2 represents a set of reference points detected on an input 3D image by image analysis with 3D image analysis software,

[0087] [Fig.3]

[0088] Figure [Fig. 3] illustrates the concept of a reference frame on a 3D image.

[0089] [Fig.4]

[0090] Figure 4 shows, from a 2D input image in front view, the identification of relevant reference points for modeling the lips according to the invention.

[0091] [Fig.5]

[0092] Figure 5 represents detail B of Figure 2.

[0093] [Fig.6]

[0094] Figure 6 represents detail A of Figure 5.

[0095] [Fig.7]

[0096] Figure 7 is a table showing the dimensional parameters of the lips extracted from the input 2D image and those from the input 3D image.

[0097] [Fig. 8]

[0098] Figure 8 represents the parameters of Figure 7 calculated from the 2D image analysis software,

[0099] [Fig.9]

[0100] Fig.9 represents the parameters of Fig.7 calculated from the input 3D image analysis software.

[0101] [Fig. 10]

[0102] Fig. 10 illustrates one method of implementing the manufacturing process for a customized applicator according to the invention.

[0103] Figure 1 shows the identification of landmarks on a 2D input image of the face, in frontal view. The 2D input image is represented by a collection of 2D landmarks characterizing the surface to be analyzed, obtained by capturing a 2D image, notably from a mobile phone. The landmarks are represented by circles (not identified by a number), the circles being larger for landmarks located at the corners of the lips and eyes, which can also be detected on the 3D input image.

[0104] Figure 2 shows the identification of landmarks (or nodes) on a 3D image of the face, in frontal view. The input 3D image is represented by a collection of points characterizing the surface to be analyzed, obtained by capturing a 3D image, notably from a mobile phone. The largest landmarks are located at the corners of the lips and are also those detected on the 2D input image.

[0105] Figure 3 represents a 3D reference frame with x, y and z axes according to the invention. It is used for each modeling.

[0106] In [Fig.4], the reference points of a 2D input image, used to determine the lip contour on the 3D output image, have been identified. These reference points are the reference points Xo, Xb, X2, X3, X4, X5, X6, X7, X8, X9 delimiting the lip contour B and several second reference points Y0, Yi, Y2, Y3, Y4, Y5, Y6, Y7 located on either side of the separation line A of the upper and lower lips.

[0107] Figures 5 and 6 show an enlargement of the lip area from the 3D input image of Figure 2. The key landmarks for evaluating lip depth can be seen, namely two landmarks 405 and 835. limiting the corners of the lips and four points 21, 24, 25, 27 located on the longitudinal axis X of the lips.

[0108] In [Fig. 7], a table shows the dimensional parameters extracted from the input 2D image and the dimensional parameters extracted from the input 3D image. These parameters can be extracted using the software indicated below:

[0109] - For extracting the dimensional parameters of the represented 2D input image In [Fig.4], the following software can be used: Dlib, Modiface.

[0110] - For extracting the dimensional parameters of the input 3D image represented In [Fig.5], the following software can be used: Truedepth from Apple, Structured light, FaceMesh.

[0111] As can be seen in this table, the dimensions describing the lip contour, obtained by the 2D input image analysis software, are L, Hiup, Harc, L^, Hiow, H2up, L2, Hc, Hiow3, L3. The physical meaning of each dimension is shown in [Fig. 8]. The dimensions describing the lip depth are L, Pmid, Psup, Pinf, their physical meaning being shown in [Fig. 9]. Note that Pmid is defined as being equal to 25% of L, L being defined in [Fig. 8].

[0112] Figure 10 illustrates an implementation of a method for manufacturing a personalized applicator according to the invention. The method comprises capturing an image by selfie / scan, then 2D processing 110 and 3D processing 111 of the captured image, to produce 112 a personalized 3D printable applicator.

[0113] The 2D 110 processing is carried out using a Dlib algorithm, namely an open source library of tools enabling facial detection of sixty-eight x, y coordinates of the face or a Modiface algorithm.

[0114] The 3D processing 111 is carried out using an Apple truedepth algorithm and ARkit, namely a library offered by Apple enabling facial recognition using the 3D Truedepth capture technology.

[0115] The 2D processing 110 leads to the determination 113 of a plurality of points on the lip contour. The 3D processing 111 leads to the detection 114 of lip landmarks and the determination 115 of lip depth.

[0116] A detection 117 of at least two points common to the 2D image and the 3D image, in particular to determine a dimensional scale, is carried out. The determination of the lip depth with the detection of landmark points of the lip contour and the dimensional scale leads to the estimation of a 3D model of the lips and to the determination 121 of the lip dimensions.

[0117] Of course, the invention is not limited to the embodiments just described.

Claims

Demands

1. A method for manufacturing a custom applicator for applying a cosmetic composition to the lips, this applicator having an application surface made of a material capable of absorbing the composition, the method comprises the following steps: i. Capturing a 2D input image equipped with a first dimensional reference frame, of at least a portion of the lip surface, ii. Capturing a 3D input image equipped with a second dimensional reference frame, of at least a portion of the lip surface, and iii.To produce (112), from at least the input 2D image, equipped with the first dimensional reference frame, and the input 3D image, equipped with the second dimensional reference frame, at least a part of the applicator or of a mold used for its manufacture, by machining a preform or by additive manufacturing, characterized by the fact that it includes the determination (117) of at least one reference point visible both on the input 2D image and on the input 3D image and the assignment to this reference point of a dimensional coordinate on an output 3D image.

2. A method according to claim 1, characterized in that it comprises the determination (113) of a plurality of points of the lip contour, from the input 2D image, and the estimation of the lip contour on the output 3D image, by interpolation from these points.

3. A method according to the preceding claim, characterized in that it comprises the determination (115) of the lip depth on an output 3D image from the input 3D image.

4. A method according to any one of the preceding claims, characterized in that it comprises the detection, on the input 2D image, of several first reference points (Xo, Xb X2, X3, X4, X5, X6, X7, X8, X9) delimiting the contour (B) of the lips and of several second reference points (Yo Yi, Y2, Y3, Y4, Y5, Y6, Y7) located on either side of the separation line (A) of the lips, in order to produce the contour of the lips on an output 3D image.

5. A method according to any one of the preceding claims, characterized in that it comprises the detection (114), on the input 3D image, of several third reference points (405, 835) delimiting the corners of the lips and of several fourth reference points (21, 24, 25, 27) located on the longitudinal axis (X) of the lips, in order to produce (121) the depth of the lips on an output 3D image.

6. A method according to any one of the preceding claims, characterized in that it comprises the display of a printable and / or manufacturable output 3D image.

7. A method according to any one of the preceding claims, characterized in that it includes positioning information for a mobile image capture device relative to the lip area, in particular by using a position sensor.

8. A method according to any one of the preceding claims, characterized in that it comprises the generation of a reworked 3D output surface, in particular by stretching the 2D input image, the applicator or mold used for its manufacture having a shape given at least in part by this reworked surface.

9. A method for lip makeup, comprising the application of a cosmetic composition to the lips using an applicator obtained by the method according to any one of the preceding claims.

10. A method for computer modeling at least one area of ​​the lips, the method comprising the operations of: i. Capturing a 2D input image equipped with a first dimensional reference frame, of at least a portion of the lip surface, ii. Capturing a 3D input image equipped with a second dimensional reference frame, of at least a portion of the lip surface, and iii. Generating a 3D output image of the lip area from the 2D input image and the 3D input image, the lip contour, iv. Determining at least one reference point visible on both the 2D input image and the 3D input image and assigning to this reference point a dimensional coordinate on the 3D output image.

11. A method according to the preceding claim, characterized in that it comprises determining the lip contour of the output 3D image from the input 2D image.

12. A method according to claim 11 or 12, characterized in that it comprises determining the lip depth of the output 3D image from the input 3D image.

13. A computer 3D modeling system for at least one area of ​​the lips, preferably intended for use in the manufacture of a personalized applicator for the application of a cosmetic product to the lips, characterized in that the system comprises at least one mobile 2D and 3D image capture device, in particular a smartphone, a system in which, once the mobile image capture device is placed in a predetermined position relative to the lips, the mobile image capture device is capable of capturing an input 2D image with a first dimensional reference frame, of at least a part of the surface of the lips, and of capturing an input 3D image with a second dimensional reference frame, of at least a part of the surface of the lips, the system further comprising a processor capable of generating an output 3D image of the lip area from the input 2D image and the input 3D image,by determining the lip contour of the output 3D image from the input 2D image and determining at least one reference point visible on both the input 2D image and the input 3D image and assigning this reference point a dimensional coordinate on the output 3D image.

14. A personalized applicator for applying a cosmetic composition to the lips, characterized in that it is capable of being obtained by the process according to any one of claims 1 to 8.