Classifications

• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; 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

Definitions

• the present invention relates to a computer system for automatically simulating an aesthetic image of a real objective, for example of a face, by recognizing its characteristic features.
• the present invention therefore relates to a method and a device making it possible to carry out automatically by computer means, the simulation of an initial objective, a face for example.
• the computer system implemented is installed in a cabin or console. It essentially comprises a computer with a hard disk of at least 420 MB, a "CD ROM” reader, a digital camera and an expert system, provided with a database and powered by an interference engine, that is to say by a powerful deductive system independent of the database.
• This expert system used is of class zero plus (0+).
• the present invention therefore relates to a method and a device making it possible to simulate, by an automatic process, an ideal image, from a real image of a lens, a face for example, characterized in that the real image captured by a device digital photographic photography in color is segmented, then in that we determine the fine characteristic features of this image by a recognition of shapes, textures and outline and finally in that we modify some of these features according to criteria decision of an expert system with a knowledge base enriched by several thousand combinations and harmonies. Capturing the color image is obtained in a few seconds by means of a digital camera of the ION type or video camera with his cane digital acquisition, under appropriate illumination from light spots. It is from this image, segmented beforehand, that the method and the device according to the invention will enable the simulation of the ideal aesthetic image.
• the process essentially comprises three phases:
• An acquisition phase we analyze and recognize the shapes and contours of a digitized photographic image using techniques characterized by the use of deductive logic,.
• a restitution phase offering the desired aesthetic simulation on a large screen and / or on paper, in particular A4 format.
• the image processing software will coordinate several successive operations.
• the histograms of the 256 levels of colors or of the 256 levels of gray equivalent corresponding to each elementary segment of the image are considered.
• it determines the points where the histograms undergo a discontinuity or else pass through minima or maxima so as to optimize the method we will determine the derivatives of the variance, the mean and the coefficient of variation of the gray levels of each image line.
• contour following automata are used. Objects can thus be easily isolated from the background. The regions of interest of the image are thus delimited. In the present case, this binarization can be carried out automatically by the search for the minima of the gray level histograms in preferred directions.
• This first phase finally includes a determination of the structural effects during which a more general evaluation is made of the degree of homeomorphism of the tissues, the degree of symmetry of the image and more particularly the degree of orthogonal symmetry when the object examined is a face (chilarity). In all of these treatments, the image quality can be improved by eliminating or attenuating noise.
• the noises correspond to random fluctuations and to parasitic information caused in particular by lighting.
• noises are in the form of pixels isolated from the neighborhood. In addition, these isolated pixels are not partially correlated.
• transformations can be contextual, that is to say that the parameters of the operators are likely to vary according to an environment or a position relative to the frame of the image or compared to characteristic shapes previously detected. . We reduce the importance of these noises by applying filters to the image.
• Filters should be classified into two categories:
• the differentiation operator filters likely, on the contrary, to enhance the contrast.
• these filters are applied locally or globally on the image and they must possibly take into account the local context of the image by taking into account its statistical properties.
• these filters are intended for operations performed in binary in grayscale, or on the contrary to operate in 256 colors. In all cases, the application of these filters requires efficient management of the intermediate "caches".
• the process determines which sex the face belongs to, whether it is of the child, adolescent, adult or old man type, then brings improvements of an aesthetic nature, depending on its program. of care. Some features of the initial face are improved. So, for example, small ⁇ des, dark circles, crow's feet, freckles, acne and rosacea are greatly reduced. More generally, the characteristic features of the face are modified according to decisions by the base of the expert system.
• the software analyzes the dominant colors that appear in the different regions of the face (for example, lip color, eyes, dominant skin tone, etc.) and provides if necessary cosmetic improvements
• the data necessary for inferences from the expert system may be incomplete or uncertain
• the use of fuzzy logic makes it possible to balance the premises of the rules and the rules themselves so as to be able to infer even presence of incomplete or imprecise data Cosmetic choices may be made even if we do not have all the necessary data with precision
• this aesthetic and made-up image will be presented as the simulation of the isolated image.
• the restitution phase is reflected first by the display, in 256 colors, on a high-definition color screen of the image gradually transformed, then by the printing of a photograph in 256 gray levels thanks to an image jet ink or laser, in A4 format, and by the edinon of the "care and beauty or beauty passport" information sheet
• a print of a second photograph in 256 colors in A4 format is edited a few minutes later thanks to a thermal sublimation, inkjet or laser printer, incorporated into 1 installation, thus revealing its new simulated face on the subject examined.
• the simulated image and the corresponding information are recorded by means of a digital optical card reader reader which is part of the installation.
• the treatment sheet and the final color photograph are edited in front of the client.
• the expert system is able to simulate on request makeup styles adapted to the circumstances. We can thus make a makeup best suited to a party, an evening, either at work or even at work. stay by the sea, etc.
• the expert system is able to simulate on demand and verify the effects of a skin care program.
• FIG. 1 represents the image of a woman's face to be modified
• FIG. 2 represents a photograph of the same face in computer form in 256 gray levels or 256 colors
• FIG. 3 shows in dark gray, the portions of the image subjected to a first treatment
• FIG. 4 shows the face modified by the method of the present invention, registered under the name MAGIC LUMIERE.
• FIG. 1 is the image of the woman's face to be modified. In order to be able to carry out computer processing of the image, this image is replaced by another image obtained by means of a digital camera in 256 colors.
• FIG. 2 represents the image of this woman's face thus obtained in computer form, formed of elementary segments, each characterized by a histogram specifying for this elementary segment the level of each of the 256 colors. It is then possible to analyze and recognize the shapes of this face.
• points such as (V; Hl) are determined where the histograms undergo a discontinuity corresponding to the limit of the dark hair of the face studied. We will also reveal all the points such as (V I; H2) or (V I; H5) where the histograms pass through minima.
• the software will then analyze the dominant colors that have been highlighted on certain parts of the face. In this case, the software analyzes the color of the eyes, eyelashes, eyelids, eyebrows, then the coloring of the lips. These extracts are then modified according to the decisions of the expert system.
• Figure 3 illustrates these operations. Having defined the elements to modify (eyelids, eyelashes, eyebrows), we proceed to extract the contours of these elements. So that these appear in dark gray or white on the image. We therefore store all the pixels in these zones in order to modify their color and achieve the desired makeup.
• FIG. 4 corresponds to the restitution phase during which the fine characteristic features of the initial image have been transformed to give an aesthetic and made-up image.
• This process and device can obviously be applied to the simulation of an aesthetic image of a face, from a simple black and white or color photograph, making it possible to perform, from the scanning of this photograph by a scanner high definition, photographic portraits of a subject by improving the aesthetics of this face by capillary and make-up techniques, blurring or rectification of certain anatomical defects, change in skin color (tanning), in this
• the database corresponding to the aesthetic simulation of the face is completed by a collection of aesthetic rules concerning makeup, hairstyling, hair techniques, skin texture and / or plastic surgery, thus making it possible to treat simultaneously the face and the hair part of a subject to edit and restore a new rewarding and ideal image
• the expert system database is enriched with rules concerning the anatomy and the corresponding surgical technique. After healing, this last image must be compared with the result obtained in order to see if there is still reason to make some additional corrections.
• the image acquisition and processing system associated with an expert system makes it possible to carry out a diagnostic aid model and to suggest the nature of the treatment to be performed (ointment, etc.).
• the device makes it easier for the dermatologist to appreciate the evolution of the condition and to easily observe the changes in color of the skin tissue (eczema, shingles, psoriasis, dermatitis, changes in color of the skin spots likely to cancer, nevi, evolving scars, bedsores, vasomotor disorders, mucosal disorders, etc.).
• the database of the expert system is enriched with information concerning the changes in color of the skin tissue or skin spots corresponding to a set of conditions, accompanied by examples of the treatments, making it possible to suggest the nature of the treatment to be carried out.
• the device according to the invention helps to make the choice of a better aesthetic by facilitating the superposition of different types of hairstyle and different hair colors on a visase.
• This application can be implemented directly with the MAGIC LUMIERE system, the application to the hair technique can be directly included in the facial treatment.
• the database corresponding to the aesthetic simulation of the face is completed by a collection of esthetic rules concerning the hairstyle and the hair color thus making it possible to simultaneously treat the face and the capillary part of a subject.
• the device according to the invention allows the eyewear optician to help make the choice of a better aesthetic by facilitating the superposition of different spectacle frames or different colored lenses on a face.
• This application can be implemented directly with the MAGIC LIGHT system, the application to the technique of glasses or optical instruments can be directly included in the treatment of the face.
• the database corresponding to the aesthetic simulation of the face is completed by a collection of aesthetic rules concerning spectacle frames and optical instruments (colored lenses, etc.) thus making it possible to simultaneously treat the face and superimpose different glasses frames adequately or different colored lenses on the face.
• the system according to the invention making it possible to recognize the face of a known person without fail can be used to control an opening system.
• the database is filled with information corresponding to all the characteristic features of one or more known faces making it possible to recognize the face of a so-called self-assertive person for sure.
• the system according to the invention also makes it possible to recognize a face partially masked or totally masked by a fabric and to compare it to a database of features characteristic of a face, and this whatever its orientation
• a 3D reconstruction system from ' 2D images whose orientation is unknown allowing the reconstruction of the facies and its face and profile representation for comparison with a file.
• the database is enriched with information corresponding to all the characteristic features of faces making it possible to reconstruct a facies from the front and from the profile.
• Another database is associated with this database. data containing faces of people which will allow their comparison with the reconstructed facies from the front and from the side.
• the system according to the invention makes it possible, during a photograph taken in a digital photography booth, to improve the aesthetics and enhance the image of a subject.
• the database corresponding to the aesthetic simulation of the vase is completed by a collection of aesthetic rules concerning make-up, hairstyling, hair techniques and skin texture, thus making it possible to simultaneously treat the face, the capillary part of a subject. to edit photographs that are rewarding for the client.
• the system according to the invention makes it possible, from a photograph of the face of a subject, supplied on conventional photographic support in black and white or in color, after having been scanned by a digitizing apparatus, to be retouched and aesthetically improved.
• the database corresponding to the aesthetic simulation of the face is completed by a collection of aesthetic rules concerning make-up, hairstyling, hair techniques and skin texture, thus making it possible to simultaneously treat the face, the capillary part of a subject for editing rewarding portraits.
• the device according to the invention makes it possible to observe and "read” with meticulousness a face the wrinkles and fine lines, the balance between the upper and lower part, the size and the shape of the nose, the earlobe and the commissures lips, etc., so it is possible to detect and determine the personality of others.
• This application can be implemented directly with the MAGIC LUMIERE system.
• the system database is enriched with a set of techniques corresponding to the observation of faces, supplemented by those of psychologists, psychiatrists and anatomists in this field.

Landscapes

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

Priority Applications (1)

Applications Claiming Priority (2)

Publications (1)

Family

ID=9470432

Family Applications (1)

Country Status (3)

Cited By (11)

Families Citing this family (5)

Citations (4)

• 1994
  • 1994-12-29 FR FR9415894A patent/FR2728982A1/fr active Granted
• 1995
  • 1995-12-28 AU AU44522/96A patent/AU4452296A/en not_active Abandoned
  • 1995-12-28 WO PCT/FR1995/001747 patent/WO1996021201A1/fr active Application Filing

Patent Citations (4)

Non-Patent Citations (3)

Also Published As

Similar Documents

Legal Events