Classifications

• • G—PHYSICS
  • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
  • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
  • G09G5/00—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators
  • G09G5/02—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators characterised by the way in which colour is displayed
  • G09G5/026—Control of mixing and/or overlay of colours in general
• • A—HUMAN NECESSITIES
  • A45—HAND OR TRAVELLING ARTICLES
  • A45D—HAIRDRESSING OR SHAVING EQUIPMENT; EQUIPMENT FOR COSMETICS OR COSMETIC TREATMENTS, e.g. FOR MANICURING OR PEDICURING
  • A45D44/00—Other cosmetic or toiletry articles, e.g. for hairdressers' rooms
  • A45D44/005—Other cosmetic or toiletry articles, e.g. for hairdressers' rooms for selecting or displaying personal cosmetic colours or hairstyle
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T11/00—2D [Two Dimensional] image generation
• • G—PHYSICS
  • G06—COMPUTING; CALCULATING OR COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T11/00—2D [Two Dimensional] image generation
  • G06T11/001—Texturing; Colouring; Generation of texture or colour
• • G—PHYSICS
  • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
  • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
  • G09G2340/00—Aspects of display data processing
  • G09G2340/06—Colour space transformation
• • G—PHYSICS
  • G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
  • G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
  • G09G2340/00—Aspects of display data processing
  • G09G2340/12—Overlay of images, i.e. displayed pixel being the result of switching between the corresponding input pixels

Definitions

• the present invention relates to a method for simulating a hair with variable colorimetry and a device for implementing said method.
• Simulation methods are known today that allow a person to visualize themselves with a new hairstyle before the hairstylist actually makes the cut and / or the color on the hair.
• the patent application EP 2 124 185 thus describes a method for finely detecting the outline of a hairstyle in order to propose to a person a simulation of his face with another hairstyle.
• the application WO 98/21695 describes a hairstyle simulation method in which a hairstyle can be chosen in a database by a person, then geometric transformations are applied to the hairstyle chosen to obtain a hairstyle that can be superimposed. likely to an initial photo of the person. According to this method, it is also possible to change the color of the hairstyle to make it consistent with the color of the initial hairstyle of the person, to further increase the likelihood.
• the present application describes a method of simulation of a hair colorimetric variable to simulate, in real time, in detail and with a reasonable technical cost at a consumer point of sale the new color that will have the hair of a person after application of a lotion of his choice, according to his current or future hairstyle.
• the invention relates to a method for simulating a hair with variable colorimetry comprising:
• the preliminary recording in a first database of a given number of images of hairstyles of given styles, and for each hairstyle image of the same style, of at least one associated colorimetric conversion matrix making it possible to obtain said hairstyle image by applying said die to a reference hairstyle image of the same style;
• identifying in the first database a reference hairstyle image of the same style as that of the hairstyle-like image and a colorimetric conversion matrix associated with the hairstyle-like image the creation from of the initial digital image of the person of an avatar superimposed on the image of the face of the person previously extracted from said initial image, a hairstyle image reconstituted by application to a reference hairstyle image of said first database a colorimetric conversion matrix associated with the hairstyle-like image;
• each hairstyle has a given hairstyle and hair type
• said associated colorimetric conversion matrix makes it possible to convert the colorimetric value of each pixel of said hairstyle image with respect to the hairstyle image. colorimetric value of the corresponding pixel of a reference hairstyle image of the same hairstyle.
• said conversion matrix is a square matrix and the colorimetric value of a pixel its component according to the colors Red, Green, Blue (RGB component).
• the colorimetric conversion matrix by lotion makes it possible to convert the colorimetric value of each pixel of a reference hairdressing image with the colorimetric value of each pixel of the same reference hairdressing image after application of said lotion on a given hair of the type of hair from which comes the reference hairstyle.
• the step of creating the avatar can be automatic or semi-automatic.
• the reconstructed hairstyle image is obtained by applying the conversion matrix associated with the hairstyle image resembling a reference hairstyle image of the database of the same style as the hairstyle-like image.
• the image of reconstituted hairstyle is then faithful in style and color to the image of hairstyle of the person.
• the reconstituted hairstyle image is obtained by applying the conversion matrix associated with the hairstyle image resembling a reference hairstyle image of said database of a style different from that of said image. like hair style.
• This variant can be implemented for example when the person wants to visualize with a different style of hairdressing.
• the method according to the first aspect further comprises, for a given reference hairstyle, the prior registration in said first database of a given number of reference hairstyle images according to various shooting parameters. and recording one or more conversion matrices each linked to one of said parameters and making it possible to convert the colorimetric values of each of the pixels of a reference image associated with said parameter from those of each of the pixels of an image reference defined as the main reference image. It is thus possible to take into account the shooting parameters of the image of the person at the point of sale and to possibly correct the shooting parameters of the reference images.
• the method according to the first aspect further comprises, during the prior registration of said first database and / or said second database, a step of simplifying the colorimetric conversion matrices comprising the definition of matrices. colorimetric conversion by areas of the image.
• This simplification step makes it possible to adapt the smoothness of the simulation to the calculation capacities available at the point of sale and thus to find the best compromise on the quality of the simulation.
• the colorimetric conversion matrices of said hairstyle images each form part of a set of metadata associated with each of said hairstyle images.
• the metadata, multi-format may further include attributes of the hairstyle image, including for example shooting parameters. Saving as metadata reduces the size of databases.
• the invention relates to a device for simulating a hair of variable colorimetry comprising:
• Storage means comprising:
• a first database with a given number of images of hairstyles of given styles, and for each hairstyle image of given style, at least one associated colorimetric conversion matrix for obtaining said hairstyle image by application of said matrix to a reference hairstyle image of the same style;
• a second database with, for a given number of hair lotions, at least one colorimetric conversion matrix per lotion; means for acquiring an initial digital image of a person; a calculation unit allowing:
• said storage means are arranged in a server remote from the computing unit, and further comprising a remote connection between said server and the computing unit.
• Figure 1 is a block diagram describing the step of building a database of hair and their colorimetric variants, according to an exemplary embodiment
• Figure 2 is a block diagram describing the step of building a database of lotions and their effects on the hair, according to an exemplary embodiment
• FIG. 3 a block diagram describing an exemplary embodiment of a method for simplifying the colorimetric conversion and / or transformation matrices, with a view to constituting the database of the colorimetric metadata of the reference hairstyles as well as the colorimetric metadata database of hair lotions.
• Figure 4 is a block diagram describing the step of simulating the hair of a person and the creation of personal color metadata, according to an exemplary embodiment
• Figure 5 is a block diagram describing the step of simulating the effect of a lotion on the hair of a person, according to an exemplary embodiment
• Figure 6 a diagram describing an embodiment of a device for implementing the method according to the invention.
• Figures 1 to 5 describe the different steps of the implementation of the method according to the invention according to an exemplary embodiment.
• FIG. 1 illustrates a first step comprising the creation of a first database 101, associating with a given number of hair styles and hair colors, colorimetric conversion matrices. It is thus necessary to carry out beforehand a cartography of a large number of styles of hairstyle and colors of hair. This preliminary barn for creating the first database is for example carried out in a specialized laboratory.
• a first step S 10 high resolution digital images of different styles of hairstyles (A, B) belonging to different people (100A, 100B) are acquired.
• style of hair By style of hair, one generally hears the arrangement of the hair (hair not attached, short, with fringe, tied in a bun, with braids, etc.) as well as possibly its texture (stiffness, suppleness, shine, friezes, etc.).
• Each image is a reference image for the hair style.
• These images are taken for a given color of the hair, for example a light color, for example blond, which will be the reference color.
• the reference images of the different hairdressing styles are taken under standard shooting conditions, these conditions being in particular the lighting, the shooting distance, the angle of view, etc.
• several images can be taken in different shooting conditions. One can then define among these reference images a main reference image for the style of hairdressing.
• a set of data is extracted from these images, in addition to the colorimetric value of each pixel of the image, parameters related to the shooting conditions and descriptive parameters of the image (for example its resolution).
• all of these data forming attributes of the image recorded as metadata associated with each image these metadata can be of multimedia form.
• metadata may include text, images, texture formulas, photographs, and so on.
• the metadata of the different reference images may also comprise one or more conversion matrices related to each of the shooting parameters. , and for converting the color values of each of the pixels of the reference image from those of each of the pixels of the image defined as the main reference image.
• the ChAo images form a set of reference images taken with different shooting conditions of the person 100A presenting the hairstyle A
• the ChBo images form a set. reference images taken with different shooting conditions of the person 100B presenting the hairdressing style B.
• an image will be defined for each style of hairdressing, A or B, which will form a main reference image of the considered hairstyle. This step can be reproduced for a number N of hair styles.
• a hairstyle of a given hairstyle for example A, but with different types of hair, for example various natural colors are made.
• This step is reproduced for each of the hair styles.
• M images each corresponding to a color, of the order for example of ten or a hundred natural colors according to the objective referred to in the exemplary embodiment concerned indicated respectively ChAi, ChA 2 , etc. for the style of hairdressing A and ChBi, ChB2, etc. for the hair style B in Figure 1.
• a high-resolution digital image of a person with a given hair style and a given type of hair can be acquired, then images of the same hairstyle but different types of hair can be calculated from the image acquired to constitute the images 111A - 11 A.
• a step S 13 then allows for each hair style and each natural color given, to extract a colorimetric conversion matrix corresponding to a given natural color.
• the conversion matrix makes it possible to convert the colorimetric value of each pixel of the hairstyle image (for example ChAi) from the colorimetric value of the corresponding pixel of the image of the hairstyle of reference of the same hairstyle ( or main reference hairstyle), for example ChAo. It is possible in some cases that a geometric transformation is necessary between two images of hairstyles of the same style, the images being acquired on two different people. In this case, the correspondence of the pixels from one image to the other is understood after application of said geometric transformation.
• each image is also associated with metadata that can include in addition to the colorimetric value of each pixel of the image, the colorimetric conversion matrix and optionally, parameters related to the shooting conditions as well as descriptive parameters of the image. 'picture.
• a first database 101 is obtained in which are stored a set of metadata each associated with a hairstyle image and notably comprising the associated colorimetric conversion matrix (s).
• the database may be enriched as a function of time of new images corresponding to new styles of hairstyles and / or associated natural colors, with each time an update of the metadata associated with the different images, and in particular an update.
• day colorimetric conversion matrices may be enriched as a function of time of new images corresponding to new styles of hairstyles and / or associated natural colors, with each time an update of the metadata associated with the different images, and in particular an update.
• Figure 2 illustrates a second step of creating a second database 102 associating a given number of hair lotions with at least one colorimetric conversion matrix by hair lotion.
• the colorimetric conversion matrix of a lotion makes it possible, for example, to convert the colorimetric value of each pixel of a reference hairstyle with the colorimetric value of each pixel of the same hairstyle after application of said hair lotion on a type of hair from which the reference hairstyle is obtained.
• a first step S20 different hair lotions are applied to different persons with various natural hair styles and colors and high-resolution digital images of the hairstyles are obtained after application of the lotion.
• a Li hair lotion is applied to different people 111A, H2A, etc. presenting the hair style A and the hair types ChAi, ChA 2 , respectively and a lotion L 2 is applied to the person 114 B presenting the hair style B with the hair type CI1B4.
• High resolution images are taken from the hairstyles once the lotion is applied to the hair and the hairstyle redone according to the initial style (121A, 1 2A for the application of the Li lotion to the hair ChAi and ChA 2 and 124B for the application from the lotion L 2 to the hair CI1B 4 ).
• the initial style 121A, 1 2A for the application of the Li lotion to the hair ChAi and ChA 2 and 124B for the application from the lotion L 2 to the hair CI1B 4 .
• step S21 for each image resulting from the application of the lotion, metadata associated with the images are extracted, then colorimetric conversion matrices by lotion are calculated and recorded (step S22) allowing the conversion of the colorimetric value of each pixel. of the image of the hairstyle after applying the lotion (for example 121A) from the colorimetric value of the corresponding pixel of the reference hairstyle of the same hairstyle (or main reference hairstyle), for example ChAo.
• This calculation and recording step can be performed for different hair lotions applied to different styles of hairdressing according to a method equivalent to that described in step S20 above, as shown in FIG. 2.
• a statistical simplification of the transformation matrices obtained making it possible to define for each hair lotion a single simplified transformation matrix, or a limited number of transformation matrices.
• the transformation matrix associated with a given hair lotion is similar to a set of styles and / or a set of hair colors, to limit, by lotion, the number of transformation matrices.
• the colorimetric conversion matrices are square matrices also called "Look Up Tables" or LUT according to the English expression, to give the component RGB (abbreviation of the English expression "Red Green Blue” ) of each pixel of the final image according to the RGB component of the corresponding pixel of the reference image. Each pixel is therefore associated with a colorimetric conversion matrix.
• the simplifications aim to simplify the matrices by defining areas of the image, or masks, in which the colorimetric characteristics of the image are sufficiently homogeneous to be able to define identical colorimetric conversion matrices for all the pixels of the zone.
• the simplifications also aim at simplifying the conversion matrices by defining sufficiently homogeneous zones to be able to define diagonal conversion matrices, or simplified compared to the initial matrices.
• the masks thus defined are part of the attributes of the image and the metadata associated with the images may include succession of couples (zone / LUT) that can be recorded in the databases.
• FIG. 3 illustrates, in a particular embodiment of the invention, a method for simplifying LUTs between an initial reference image and a target image 300 whose conversion matrix is recorded in the database.
• a first step S31 consists in selecting a given number of zones Zi estimated homogeneous on the target image 300, this number in front of one side being as small as possible to facilitate the subsequent processing of the image but the largest possible (that is to say with very small areas, the extreme being a zone per pixel) if we aim that each simplified matrix obtained is as close as possible to the real pixel-by-pixel transformation matrix. There is therefore an optimization to be found in the simplification which will be a function of the final cost of treatment sought at the point of sale.
• Each iteration consists in selecting a number N of zones not necessarily disjointes, compute (step S32) the simplified LUTs of each zone then calculate (step S33) for the entire image a function called "level of accuracy".
• This function consists in calculating the difference between on the one hand the image 301 obtained using the application of the simplified LUTs of all the zones with the initial reference image and on the other hand the image 302 of the initial unimproved LUT transforming pixel by pixel of the initial reference image.
• step S31 The methods of selecting and selecting a certain number of zones (step S31) can be very diverse. Also image editing professionals commonly use area definition tools, either by means of geometric contour selection or by methods related to color histograms; they then realize the transformations they want on these areas. For example, the aforementioned patent application WO98 / 21695 uses a method of selecting zones by histograms.
• the parameters LIFT, GAIN, GAMMA of said linearity are calculated by RGB color channel using a statistical method of least cost of dispersion (for example the variance).
• step S33 a statistical computation of the dispersion of the entire image between the reality and the resultant resulting from the simplified LUTs is performed (step S33).
• the result of this calculation is called the "precision level" of this first selection of zones.
• step S34 The above process is then repeated (step S34) by selecting a different number of zones (either smaller or larger) and comparing the precision levels found at the end of the process.
• the operator will eventually find (step S35) the characteristics best suited to the hairstyle for which the simplification has been made and the model of use at the point of sale with the corresponding level of precision required. It is these characteristics (Zi zones / associated conversion matrices) that will then be recorded in the color metadata database 101 that is to be constituted (step S36).
• a similar simplification process can be applied to the conversion matrices for creating colorimetric metadata associated with hair lotions in the second database 102.
• a retouching professional can work on the image of the reference hairstyle by successive operations thanks to the tool at his disposal (for example Photoshop®), in order to get as close as possible to using of these operations the target image of the hairstyle to be reproduced via colorimetric transformations on the image of the reference hairstyle.
• Operations with all their characteristics, including the selected areas and the associated LIFT, GAIN, GAMMA coefficients, are then listed in a metadata list, a "color decision list" describing a sequence of color metadata that is then exported to the database associated with the target image.
• the level of precision obtained by the manual operations thus carried out by the professional is also calculated.
• mathematical tools for massive data categorization such as those used for credit or financial market risk assessments, are used to perform almost all of the above operations almost automatically, including part successive reiterations.
• the zones are then chosen according to a geometric or colorimetric algorithm in a systematic manner and according to certain criteria, for example criteria of geometric shapes with increasing surface or criteria for selecting histograms of gray shades (see patent application WO98 / 21695 previously cited).
• a ranking algorithm uses the "level of precision" criterion and / or any other criteria considered relevant, for example the number of electronic operations necessary for the treatment at the point of sale. this allows to identify the best compromise between number of zones and accuracy sought and then save it in the database.
• FIG. 4 illustrates a third step of the method according to the invention implemented this time directly on the place where the simulation of the colorimetry of the hair of a person 400, called in the rest of the description, must be carried out.
• a first step S40 acquires a digital image 400 of the person.
• This image can be recorded with light technical means, accessible at the point of sale, general public or semi-professional (digital camera, computer tools, smart phones or tablets) and is not necessarily a high resolution image.
• a step S41 the hair 401 and the face 402 are cut off without the hair, starting from the image 400.
• the extraction of the hair 401 can be made by known means, for example those described in the patent application. EP2124185 previously mentioned.
• the first database 101 is searched for the hairstyle closest to that photographed, both in hair style and in color.
• This step may comprise the treatment of the initial image 400 of the hairstyle, by applying a pre-established model associated with a retouching tool in order to characterize it an approximate first time with a few simple parameters predefined by the system (shapes, main colors, identified textures). It is then possible to carry out the automatic extraction in said first database 101 hairstyles and hair satisfying the initial approximate simple criteria found above, by automatic comparison of resemblance of the characteristics of the hairstyle / hair photographed with those extracted, then d display on a screen the few images that correspond to the optimization of this comparison. A manual selection of the hairstyle / hair whose likeness is perceived by the user to be the best can then be performed if necessary.
• a step S43 it is then extracted in said first database metadata (zoning masks, textures, colors, shapes, etc.) associated with the image of the hairstyle / hair determined in the step previous. It is then possible (step S44) to reconstruct a hairstyle from the reference image corresponding to the identified hairstyle of the person and by application of the conversion tables. ("LUT") or couples (masks / LUTs) present in said metadata associated with the image of the hairstyle selected in the database. This hairstyle is associated with the previously registered face 402 of the person to form the reconstructed image 410 (step S45).
• LUT conversion tables
• an equipment and a software application give the person the HMI tools varying a finite number of characterization parameters of his hairstyle / hairstyle, from his initial reconstituted avatar.
• This is an independent module that can for example complement the tools of composition / calibration that can currently be found on the market.
• This module can either be sold as a complement to these tools (such as Photoshop®), or sold separately by coming to interface with these applications as a nested module ("plug-in").
• FIG. 5 illustrates a fourth step of the method according to the invention corresponding to the visual simulation of the application of a hair lotion.
• a step S 1 the colorimetric conversion matrices associated with the various hair lotions previously recorded in the second database 102 are applied, in order to simulate the effects of the lotion chosen on the avatar 410 of the person created in the step previous ( Figure 4).
• Images 501, 502 obtained after applying the matrices corresponding to one or other of the lotions L 1 , L 2 in the example of FIG. 5 can then be displayed (step S52).
• the colorimetric conversion matrix applied is that corresponding to the lotion chosen, defined for the hairstyle / type of hair identified for the person.
• the matrix can be defined by zone and in each zone, by pixel color level.
• FIG. 6 illustrates a system for implementing the invention according to an exemplary embodiment.
• It includes a server 61 accessible remotely, for example via the Internet, in which there are storage means with the elements of the first and second databases of characterization of hairstyle / hair type and hair lotions.
• said databases have been designed beforehand and can be updated in a central laboratory (not shown).
• the databases are stored separately but are connected to each other.
• the server may further include downloadable software for implementation at the point of sale of the method of simulation of a hair colorimetric variable, as described beforehand.
• It furthermore comprises, on the point of sale, a digital image acquisition device 64 and a computing unit 62 for the actual implementation of the simulation of the person's hair.
• the computing unit 62 may comprise a device for downloading the elements created or modified from the centralized laboratory to the accessible server.
• the computing unit may finally include equipment as well as software and possibly downloadable data at the point of sale of the hair lotions.

Landscapes

• Engineering & Computer Science (AREA)
• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Theoretical Computer Science (AREA)
• Computer Hardware Design (AREA)
• Image Processing (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=47177934

Family Applications (1)

Country Status (4)

Families Citing this family (14)

Family Cites Families (4)

• 2011
  • 2011-10-18 FR FR1159396A patent/FR2981254A1/fr not_active Withdrawn
• 2012
  • 2012-10-18 US US14/352,878 patent/US20140306982A1/en not_active Abandoned
  • 2012-10-18 WO PCT/EP2012/070689 patent/WO2013057210A2/fr active Application Filing
  • 2012-10-18 EP EP12784506.3A patent/EP2768339A2/de not_active Withdrawn

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events