METHOD AND DEVICE FOR DETERMINING COLOR HOMOGENEITY OF HAIR

Description

[0001] The invention relates to a method and a device for determining the color homogeneity of hair.

[0002] Uniform hair coloring can be an important cosmetic goal.

[0003] Apart from unusual fashion statements, inhomogeneous coloring, for example with spots or streaks, can be undesirable.

[0004] Furthermore, it may also be undesirable if, in an initially possibly homogeneous hair coloring, color inhomogeneities develop over time, for example caused by UV radiation, washing processes or similar.

[0005] Color homogeneity can therefore represent an important consumer-relevant parameter.

[0006] DE102014220077 A1 describes a method for identifying a hair color in an image by comparing an identified hair color distribution with reference hair color distributions. Document US2011194762 A1 describes a method for identifying hair in an image by first determining a confidence image of a head region and then analyzing this confidence image. Document XP055407443 describes a method for recognizing a head in color images by analyzing the image in a color space. Document WO2015044944 A2 describes a method for determining hair from a plurality of images of hair taken from different perspectives. Document US2012 / 075441 A1 describes a method for determining hair properties using spectroscopy. Document US 2012 / 075331 A1 describes a method for realistically changing the hair color in an image to a target hair color.

[0007] A quantitative measure may be desirable for assessing color homogeneity, particularly for comparing different colored hairstyles in terms of color homogeneity and / or for assessing how the color homogeneity of a colored hairstyle develops over time.

[0008] In various embodiments, a color homogeneity of hairstyles can represent a target parameter that can be determined by means of image processing methods with a subsequent statistical analysis of a color angle.

[0009] In various embodiments, the method for determining the color homogeneity of hair can be implemented with little or no equipment required. For example, the method can be implemented using an app on a tablet or smartphone. This can enable a user, for example, to determine the color homogeneity of their hair without professional support and without having to provide a hair sample for manipulation, e.g.his hairstyle, for example by using a smartphone or tablet to take a digital image of the hair and determining the color homogeneity by means of the smartphone / tablet, wherein the smartphone / tablet is defined in various embodiments of a parameter set in which one of the parameters is a color tone, determining and / or setting at least one hair examination region in the digital image, determining a color tone distribution in the at least one hair examination region, and determining at least one color homogeneity value based on the determined color tone distribution.

[0010] The invention is defined in the claims, wherein claim 1 defines a method according to the invention and claim 14 defines a device according to the invention. Further preferred embodiments are defined in the dependent claims 2-13.

[0011] Embodiments of the invention are illustrated in the figures and are explained in more detail below.

[0012] It shows Figur 1A a representation of an image of hair colored in an RGB space parameterized for use in a method for determining a color homogeneity of hair according to various embodiments; Figur 1B a representation of the B-channel of the image of the dyed hair from Figur 1A ; Figur 1C a graphical representation of an intensity value distribution of the B-channel of the image from Figuren 1A and 1B ; Figur 2A a representation of the image transformed into the HSV color space from Figuren 1A bis 1C in a color channel according to various embodiments; Figur 2B a graphical representation of an intensity / hue distribution of the hue channel of the image from Figur 2A ; Figur 3A a representation of an image of hair colored in an RGB space parameterized for use in a method for determining a color homogeneity of hair according to various embodiments; Figur 3B a graphical representation of an intensity / hue distribution of the hue channel of the image from Figur 3A ; Figur 4 a table comparing color values ​​in an RGB color space and an HSV color space; Figur 5 a flowchart illustrating a method for determining a color homogeneity of hair according to various embodiments; and Figur 6 a graphical representation of a data processing device for carrying out a method for determining a color homogeneity of hair according to various embodiments.

[0013] In the following detailed description, reference is made to the accompanying drawings, which form a part of this application, and in which is shown by way of illustration specific embodiments in which the invention may be practiced. In this regard, directional terminology such as "top," "bottom," "front," "back," "fore," "rear," etc., is used with reference to the orientation of the described figure(s). Since components of embodiments can be positioned in a number of different orientations, the directional terminology is for purposes of illustration and is in no way limiting. It is to be understood that other embodiments may be utilized and structural or logical changes may be made without departing from the scope of the present invention.It is understood that the features of the various exemplary embodiments described herein may be combined with one another unless specifically stated otherwise. The following detailed description is therefore not to be construed in a limiting sense, and the scope of the present invention is defined by the appended claims.

[0014] A digital image can be understood herein as a data packet that can be represented by a data processing system as a two-dimensional (planar) arrangement of image points (also referred to as pixels), for example in a coordinate system having an x-axis and a y-axis, wherein each image point has at least one image position as an x, y coordinate pair and intensity information, wherein the intensity information can be represented, for example, as the color of a pixel of a monitor or a printed dot of a printed image. The digital image can be, for example, a photograph taken with a digital camera or an individual image of a video sequence recorded with a digital camera.

[0015] A "color" can be understood here as a combination of a hue (i.e. a spectral color impression, also known as hue or, in English, hue, which can be understood as what is regarded as the "actual color"), a color intensity (i.e. how intense the color appears, e.g. compared to a neutral gray, which is also known as saturation, color saturation, colorfulness, chromaticity, color depth or, in English, saturation) and a brightness or darkness (i.e. how light or dark the color appears).

[0016] For color representation, the color can be parameterized in a color space.

[0017] A common color space from which color information (e.g., hair color information of dyed hair or hair before dyeing, also referred to as the base hair color) originates, or in which the color information is represented, must be such that a determined or represented color is independent of the medium through which the color is determined or represented (e.g., colorimeter, screen, printer, scanner, human eye, etc.). The color space can, for example, be an L*a*b* color space. Here, the hue can be parameterized, for example, using two parameters, a* and b*.

[0018] Other common color spaces can be, for example, an RGB (parameterized in red, green, blue) or CIELUV (the hue is parameterized here in a UV plane) color space, which can be better suited for an additive light color, which is used, for example, in monitors, than, for example, the L*a*b* color space, or a CMYK color space, which, based on a subtractive color model, is used in particular in a multi-color printing area and parameterizes colors in four color channels corresponding to the basic printing colors: cyan, magenta, yellow (yellow) and black (color depth, key).

[0019] To illustrate various embodiments, the RGB color space is used here for comparison, which is to be understood as representative of all color spaces in which the hue does not represent a separate parameter but is represented as a mixture of two or more parameters.

[0020] FIG. 1A shows a grayscale representation of an image of (actually blue) colored hair 102 parameterized in an RGB space for use in a method for determining a color homogeneity of hair according to various embodiments.

[0021] Even if a color representation of the dyed hair 102 is not possible here, it can also be seen from the grayscale image 100, given a correlation between a spatial distribution of gray values ​​and a hair structure (overlying areas of curls / hair strands appear light, underlying curls / hair strands and crossing areas of curls appear dark), that each pixel of the image 100 can form at least an overlay of light and dark image portions, which likely represent illuminated and shadowed areas 102H and 102S, respectively, with the color of the hair 102. The areas that are light due to illumination or dark due to shading may have nothing to do with a color inherent in the hair 102, and thus with a color inhomogeneity, or at least do not allow a color homogeneity to be determined from the combined image.

[0022] FIG. 1B shows a grayscale representation 101 of the blue channel (also called B-channel) of the image 100 of the (actually blue) colored hair 102 from Figur 1A .

[0023] When representing the blue hair in the blue channel of the RGB-parameterized image, it could be assumed that an essentially uniformly blue-colored image would result (which in the grayscale representation could correspond to a representation with a uniform gray tone).

[0024] Based on the still in FIG. 1B However, the prominent illuminated and shadowed areas 102H and 102S (i.e. the correlation of hair structure and light / dark image areas) shows that an intensity of the pixels in the blue channel varies in order to, in combination with the red and green channels, reproduce the illuminated and shadowed areas 102H and 102S bright or dark.

[0025] A large fluctuation in the intensity of the pixels in the blue channel can also be determined by FIG. 1C , which is a graphical representation of an intensity value distribution (number of pixels depending on the intensity value) of the B-channel of the image from FIG. 1A and FIG. 1B represents, be evident.

[0026] Thus, a representation in one of the RGB channels is not even suitable for estimating the color homogeneity of hair that is dyed in a color that corresponds to one of the parameterized channels, and it is understandable that for hair colors that would already have to be represented as mixtures of the color channels, such as brown or blond, the estimation of the color homogeneity is even less possible.

[0027] FIG. 2A shows, again as a grayscale representation, a color channel representation 200 of a transformed image, wherein the transformed image according to various embodiments is derived from the RGB image 100 of FIG. 1A by transforming into the hue-saturation-lightness value color space (also called hue-saturation-value (HSV) color space).

[0028] For the representations in FIG. 2A und FIG. 2B a continuous circular hue value, which is expressible as an angle greater than or equal to 0° and less than 360°, can be mapped to a discrete distribution with 256 discrete hue values, which are represented as 256 grey values ​​of different intensity in the figure in FIG. 2A are shown, be depicted.

[0029] The hue channel display 200 shows a much more uniform hue value distribution (which is displayed as a grayscale display), which on the one hand shows the spatial correlation of the illuminated / shadowed areas with the light or dark areas (which are shown in FIG. 1B still existed) no longer has (in FIG. 2A The same places are marked as in FIG. 1A and FIG. 1B , whereby the area 102S is no longer located in an above-average dark part of the image and the area 102H is no longer located in an above-average bright part of the image), and which, on the other hand, can more closely correspond to a typical homogeneity of a hair coloring (apart from a deliberate creation of highlights and / or an improper or unsuccessful coloring).

[0030] This means that by converting to HSV values ​​(or to other color spaces that define a color angle, e.g. HSB, LCh), a complete separation of the color tone information (hue) from brightness information can be achieved, and thus an isolated consideration of color homogeneity can be made possible.

[0031] FIG. 2B is a graphical representation 201 of a hue frequency distribution of the hue channel (also called hue distribution) of the image from FIG. 2A .

[0032] In FIG. 2B shows an essentially monomodal distribution 230 of the hue values ​​with a small width.

[0033] In various embodiments, the distribution of the hue values ​​can be used to determine at least one color homogeneity value.

[0034] Statistical analysis methods can be used to determine, for example, a standard deviation (abbreviated as StdDev in the figures), also called confidence interval d, a circular variance S and / or an angular deviation s of the hue distribution.

[0035] In the color spaces described in more detail below, which have the hue as a separate parameter and are used in the method for determining a color homogeneity of hair according to the invention, the hue is expressed as an angle that can be greater than or equal to 0° and less than 360°.

[0036] According to the invention, circular statistics are used for the statistical analysis of the hue distribution. Among other things, circular statistics can be used to avoid inhomogeneities at a transition from 359° to 0° (such inhomogeneities would result in a mean of 180° between 10° and 350° when applying conventional linear statistical methods, instead of the reasonable value of 0°).

[0037] Angles α i assigned to the color tones can, in various embodiments, first be converted into unit vectors in a two-dimensional plane by means of r i = cos ∝ i sin ∝ i .

[0038] To obtain an average angle ∝, the unit vectors can be averaged in various embodiments: r ¯ = 1 N ∑ i r i

[0039] From the averaged vector rthe mean angle ∝ can be determined, if necessary specifying a confidence interval, e.g. a 95% confidence interval, by converting it using the four quadrant inverse tangent function.

[0040] A vector length R = r ¯ , which can be between 0 and 1, can already represent a measure of the color homogeneity of the hair, because the greater the color homogeneity, the closer the angles assigned to the color tones can be to each other, and the longer (i.e. the closer to 1) the vector length R can be.

[0041] As a measure of the color homogeneity, a circular variance can be used in various embodiments S = 1 − R This also lies between 0 and 1, but is smaller the closer the angles are to each other.

[0042] Furthermore, in various embodiments, an angular standard deviation (also referred to as angular deviation) can be used as a measure of the color homogeneity. s = 2 1 − R be used.

[0043] Furthermore, other values ​​known in circular statistics that provide a measure of a distribution, e.g. a scatter, of the color tones can be used.

[0044] Circular statistics can be carried out using any suitable software, for example using specially developed software, e.g. an app, or using existing software packages, e.g. Oriana or the EXCEL tool EZ-Rose.

[0045] In the digital image, as in FIG. 3A As shown, according to various embodiments, a hair representation region 334, ie, a region in which hair is depicted, can be determined using known methods. The known method can, for example, comprise a masking method, such as may be common in Photoshop and other software packages.

[0046] The hair representation area 336 can comprise a plurality of pixels of a digital image 100 that depict the hair 102 and that can form a contiguous area or a surface consisting of a plurality of individual areas. A plane in which the hair area can be arranged can be determined, for example, by the x-axis and the y-axis of the digital image.

[0047] In various embodiments, at least one hair examination region 336 can be determined and / or defined in the digital image.

[0048] In the FIG. 2A und FIG. 2B In the illustrated embodiment, the hair examination area 336 comprises substantially the entire area of ​​the digital image and is rectangular in shape.

[0049] In various embodiments, the hair examination region 336 may comprise a portion of the hair display region 334.

[0050] In FIG. 3A In the embodiment shown, one of the hair examination areas 336 for which the FIG. 3B shown distribution was determined, only a (rectangular) partial area of ​​the hair representation area 334.

[0051] In various embodiments, the digital image may have additional display areas in addition to the hair display area 334, in which, for example, objects, body parts, etc., may be displayed. In various embodiments, the hair examination area 336 may be selected such that no part of the other display areas falls within the hair examination area 336.

[0052] In various embodiments, the hair examination region 336 can have any shape. For example, in addition to being rectangular (e.g., square), the hair examination region 336 can also be triangular, polygonal with a number of corners other than three or four, elliptical, round, or any other shape. For example, the hair examination region 336 can comprise the entire hair display region 334.

[0053] In various embodiments in which the hair examination region 336 comprises the entire hair representation region, the hair examination region 336 can consequently comprise a single or multi-part area, which can be examined as a single hair examination region 336 for a subsequent analysis, e.g., for determining a value for a proportion of linear regions and / or for a distribution of linear regions, even in the case of the multi-part area.

[0054] In various embodiments, the at least one hair examination region 336 may comprise a plurality of hair examination regions 336. In various embodiments, in a subsequent analysis, e.g., when determining a color homogeneity value, each of the plurality of hair examination regions 336 may be examined separately, ie, the value for the color homogeneity of the hair may be determined separately for each of the plurality of hair examination regions 336.

[0055] In various embodiments, determining the at least one hair examination region 336 may include determining the hair representation region 334 and setting the at least one hair examination region 336.

[0056] For example, defining the at least one hair examination region 336 can mean that, as described above, the entire hair representation region is defined as the hair examination region 336, and / or one or more hair examination regions 336 can be defined, for example automatically by means of the data processing device, e.g., taking into account predetermined conditions. For example, a size and / or number of the hair examination regions 336 can be specified, e.g., by a user, and the hair examination regions 336 can then be defined automatically, e.g., by means of suitable software, e.g., such that a minimum distance is maintained between the hair examination regions 336, that the hair representation region is covered as evenly as possible, or the like.

[0057] In various embodiments, the color homogeneity can be determined for colored hair and / or for uncolored hair.

[0058] In various embodiments, the color homogeneity can be determined both before and after coloring. Based on the results, in various embodiments, a relationship can be determined between the color homogeneity before and after coloring, for example, for a plurality of colorations. In various embodiments, additional parameters describing the hair to be colored can be included, for example, the degree of graying and / or previous damage to the hair.

[0059] In various embodiments, the color homogeneity can be determined as a function of time, for example, a temporal change in the color homogeneity can be determined after one or a plurality of hair washes, at certain time intervals after exposure to UV light, or the like.

[0060] In various embodiments, the hue channel may be presented, for example, to a user, for example using a display, e.g., a display.

[0061] In various embodiments, the hue channel (Hue), while masking out the other channels (Saturation, Value), can be transformed back into a color space suitable for the display (e.g. RGB for a display or CMYK for a printer), so that possibly only the hue differences contribute to a different display of different pixels, but a realistic display can result with regard to the hues and the distances between the different hues.

[0062] In various embodiments, different hues that do not correspond to the determined hues can be assigned to the various hue tones of the hue channel (in other words, to the various hue values ​​of the hue parameter) for display purposes. This can, for example, make it possible to exaggerate small hue differences. FIG. 4 shows a table 400 with a comparison of color values ​​in an RGB color space and an HSV color space.

[0063] The RGB color space can be an additive color space that simulates color perceptions by additively mixing three primary colors (red, green, and blue).

[0064] A possible conversion from RGB to HSV can be described as follows: Precondition: R , G , B ∈ [0,1] MAX : = max R G B , MIN : = min R G B H : = 0 , falls MAX = MIN ⇔ R = G = B 60 ° ⋅ 0 + G − B MAX − MIN , falls MAX = R 60 ° ⋅ 2 + B − R MAX − MIN , falls MAX = G 60 ° ⋅ 4 + R − G MAX − MIN , falls MAX = B falls H < 0 ° , dann H : = H + 360 ° S HSV : = 0 , falls MAX = 0 ⇔ R = G = B = 0 MAX − MIN MAX , sonst S HSL : = 0 , falls MAX = 0 ⇔ R = G = B = 0 0 , falls MIN = 1 ⇔ R = G = B = 0 MAX − MIN 1 − MAX + MIN − 1 , sonst V : = MAX L : = MAX + MIN 2 Postcondition: H ∈ [0°,360°], S , V , L ∈ [0,1]

[0065] These formulas may reflect some peculiarities of HSV values. For example, if R = G = B, H may be meaningless, so by definition H = 0. If R = G = B = 0, then S may be meaningless, so by definition S = 0.

[0066] In various embodiments, modified color models, such as HSL, HSB or HSI, can be used instead of HSV.

[0067] The HSL color space (also known as HLS) can have the parameters hue (also known as hue angle) H, chroma S, and luminosity L. In contrast to the HSV color space, however, it can be referenced to the gray point lying between white and black as neutral gray. The color solid can be represented as a double cone, cylinder, or six-sided prism. The hues (chromatic values) can be located on the outside, with the gray point in the center. The CIE LCh° model can be structured similarly, with luminosity L, chroma C, and the hue (hue) angle h°, which means it can correspond to a Lab color space represented in cylindrical coordinates.

[0068] The HSB and HSI models can be oriented towards the needs of colorimetry and phototechnical reproduction. Here, too, H can stand for the hue (chromatic value, hue) and S for saturation. The difference can refer to the third coordinate: on the one hand, HSB with the radiant quantity of brightness, e.g., the absolute brightness, B, and on the other hand, the HSI color model with the luminous intensity I.

[0069] FIG. 5 shows a flowchart 500 illustrating a method for determining a color homogeneity of hair according to various embodiments.

[0070] According to the invention, the method comprises the following steps: providing a digital image on which hair is depicted and which is parameterized in a color space defined by means of a parameter set, one of the parameters being a color tone (at 510), determining and / or defining at least one hair examination area in the digital image (at 520), determining a color tone distribution in the at least one hair examination area (at 530), and determining at least one color homogeneity value based on the determined color tone distribution (at 540).

[0071] FIG. 6 shows a graphic representation of a device 600 for determining a color homogeneity of hair according to various embodiments.

[0072] According to the invention, the device 600 for determining a color homogeneity of hair comprises a data processing device 660.

[0073] The data processing device 660 may, for example, comprise a computer, a tablet, a smartphone, a laptop, a smart mirror, or any other data processing device suitable for executing the method for computer-assisted hair color consultation according to various embodiments. For simplicity, the data processing device 660 is also referred to herein as computer 660. The data processing device 660 may comprise a processor 662, for example, a microprocessor.

[0074] According to the invention, the device 600 for determining a color homogeneity of hair has a display device 664.

[0075] The display device 664 can, for example, comprise a screen of a smartphone, a PC, a laptop, a smart mirror, or another device 600 for determining hair color homogeneity. The display device 664 can be used, for example, to display results of the method for determining hair color homogeneity, to request input parameters for executing the method, or the like.

[0076] The display device 664 can be connected to the data processing device 660 via a first data connection 670. The display device 664 can exchange data with the data processing device 660 via the first data connection 670. In a case where the device 600 comprises a smartphone, a tablet, or the like, the display device 664 and the first data connection 670 can be integrated into the device 600.

[0077] In various embodiments, the device 600 for determining a color homogeneity of hair may comprise a camera 666.

[0078] The camera 666 may be configured, according to various embodiments, to capture a digital image 100 of hair 102, e.g., of a user's hair.

[0079] According to various embodiments, the at least one camera 666 may comprise a digital still camera or a video camera, ie a camera 104 which may be configured to record a plurality of individual images as a time sequence.

[0080] In various embodiments, the device 600 for determining hair color homogeneity can have a second data connection 674 between the computer 660 and the camera 666. Data can be transmitted from the computer 660 to the camera 666 by means of the second data connection 674, for example, for a conventional software control of the camera 666. Furthermore, data, for example the digital image(s) captured by the camera 666, can be transmitted to the computer 660 by means of the second data connection 674. In the event that the device 600 comprises a smartphone, a tablet, or the like, the camera 666 and the second data connection 674 can be integrated into the device 600.

[0081] In various embodiments, a camera 666 can be dispensed with in the device 600 for determining a color homogeneity of hair, for example if the digital image 100 is provided to the data processing device 660 in another way, for example by means of a data transmission.

[0082] The data processing device 660 can be configured, for example using the processor 662, to process the image received from the camera 666 or otherwise using image processing software, for example to determine the hair representation area in the received image in a known manner and, as described above for various embodiments, to determine the color homogeneity of hair. In various embodiments, the image processing software can comprise an app.

[0083] In various embodiments, the data processing device 600 may include an input device 668 for providing information to the data processing device 600, for example, a keyboard, a mouse, a touch-sensitive surface of the display device 664, or the like.

[0084] The input device 668 can be connected to the data processing device 660 via a third data connection 672. The input device 668 can exchange data with the data processing device 660 via the third data connection 672. In a case where the device 600 comprises a smartphone, a tablet, or the like, the input device 668 and the third data connection 672 can be integrated into the device 600.

[0085] Further advantageous embodiments of the method emerge from the description of the device and vice versa.

Claims

1. A method for determining a color homogeneity of hair, the method comprising: providing a digital image in which hair is depicted and which is parameterized in a color space defined by means of a set of parameters, one of which parameters is a hue, wherein the hue is defined by a color angle on a color wheel; determining and / or defining at least one hair analysis region in the digital image; determining a hue distribution in the at least one hair analysis region; and determining at least one color homogeneity value based on the determined hue distribution, wherein a statistical analysis of the hue distribution is carried out using circular statistics in order to determine the at least one color homogeneity value.

2. The method according to claim 1, wherein providing the digital image comprises transforming an original digital image, which is parameterized in an original color space, into the color space, wherein the original color space is different from the color space.

3. The method according to claim 1 or 2, wherein the color space is a hue-saturation-lightness color space.

4. The method according to one of claims 1 to 3, wherein determining at least one color homogeneity value based on the determined hue distribution comprises determining a circular variance.

5. The method according to one of claims 1 to 4, wherein determining and / or defining at least one hair analysis region comprises the user defining the at least one hair analysis region.

6. The method according to claim 5, wherein the user defining the at least one hair analysis region comprises defining a position and / or shape and / or size of the hair analysis region.

7. The method according to one of claims 1 to 4, wherein determining and / or defining at least one hair analysis region comprises: determining a hair display region in which the hair is displayed; and defining at least a portion of the hair display region as the at least one hair analysis region.

8. The method according to claim 7, wherein the at least one hair analysis region comprises the entire hair display region.

9. The method according to one of claims 1 to 8, wherein the at least one hair analysis region comprises a plurality of hair analysis regions.

10. The method according to claim 9, wherein the plurality of hair analysis regions differ from one another in their center position.

11. The method according to claim 10, further comprising: relating the determined color homogeneity values to the center positions of the hair analysis regions.

12. The method according to one of claims 1 to 11, further comprising: displaying the determined result.

13. The method according to claim 12, wherein displaying the determined result comprises displaying the digital image, wherein, in the hair display region of the digital image, the hair is displayed with the hue, wherein the entire hair display region is displayed with a uniform color saturation and a uniform lightness.

14. A device for determining a color homogeneity of hair, the device comprising: a data processing device; and a display device; wherein the device is configured, preferably by means of an app, to carry out the method according to one of claims 1 to 13.