Method of acquiring a photographic portrait of an individual and device implementing the method

Abstract

A method of acquiring a photographic portrait of an individual comprises: acquiring a biometric image of the face of an individual by means of at least one digital camera (22), simultaneously with this acquisition, illuminating the face of the individual with at least two separate light sources not centered on said face, said light sources being formed by light-emitting diodes (38) managed by a PWM system. At least two acquisitions of said face are performed separately: using standard lighting, with the aim of biometric acquisition of said face by said camera (22); using differentiated lighting, i.e. generating a lighting effect on said face by means of said at least two light sources (36, 37), also acquired by said camera. These at least two acquisitions are not synchronized and are performed randomly in a short length of time. A central control and processing unit is adapted to process said acquisitions and determine whether there is a fake after processing of the acquisition using differentiated lighting.

Description

Technical Field

[0001] This invention relates to the technical field of obtaining photographs of individuals (particularly within compartmentalized facilities or installations).

[0002] In addition to optimizing the quality of photographs, this invention particularly relates to the protection of the acquisition of such photographs, especially for the purpose of using such photographs for identification or access documents. Background Technology

[0003] Devices for acquiring photographs of individual faces are widely used. Typically, they are marketed under the trade name "Photomaton," particularly in France, and are conventionally formed by compartments defined by side walls and including passageways to the interior of the compartments, which are typically enclosed by curtains or the like. The device for acquiring the photograph is integrated into one of these walls, and is conventionally formed by a photographic apparatus or camera, illuminating the individual's face while the photograph is being acquired.

[0004] One of the challenges faced by the designers of this device was optimizing the lighting for the face of the individual in question. More specifically, the goal was to avoid "flattening" the acquired rendering as much as possible and to facilitate the differentiation of different raised areas of the face. Overexposure of certain parts of the face was also avoided.

[0005] Different solutions have been developed to eliminate these drawbacks. In particular, document EP 2860582 describes a device of the type of interest that implements two complementary luminescent surfaces laterally facing the field of view of the photographic apparatus or camera and is positioned between a wall integrating the acquisition member and the object-focal plane of the acquisition apparatus. Such a device allows for uniform illumination of an individual's face, including when they are posed in an approximate manner relative to a reference position, thereby achieving high-quality photographic portraits regardless of the shape and positioning of the individual's face at the level of the reference position of the device.

[0006] Undoubtedly, the device described in this document enables a significant improvement in the quality of the resulting photographs.

[0007] However, for the more specific purpose of taking such photographs for identity or security purposes, governments and / or security services wish to eliminate any risk of deception, as currently specified under the term "anti-deception." In practice, such deception (also known as "biometric attacks") can originate, for example, from a mask prominently placed over the face of the individual in question, and also from presenting a portrait on a flat or curved support representing a print or electronic display at the focal point of the acquisition component, thus potentially generating a stolen photograph or granting access to a secure location, thereby deceiving a database listing authorized users. Such deception can also originate from digital associations of at least two different faces, resulting in the ability to impersonate various identities already used for such associations. This digital association is also referred to as "morphing."

[0008] The present invention is therefore intended to overcome this difficulty and to combat these risks of deception. Summary of the Invention

[0009] According to the first aspect, methods for obtaining photographic portraits of individuals include:

[0010] ■ Acquire biometric images of an individual's face using at least one digital camera.

[0011] ■ Simultaneously with this acquisition, the face of the individual is illuminated by at least two separate light sources not centered on the face, the light sources being formed by light-emitting diodes (LEDs) managed by a PWM (short for "pulse width modulation") system;

[0012] The operation of the PWM system and the at least one camera is managed by a central control and processing unit.

[0013] According to the present invention, the face is acquired at least twice:

[0014] ● Using standard lighting, wherein the at least two light sources are activated simultaneously to uniformly illuminate the face, thereby allowing biometric acquisition of the face using the at least one camera, i.e. for subsequent use, for example, for identity verification or access control support;

[0015] ● Use differentiated lighting, wherein the at least two light sources generate a lighting effect on the face, and the lighting effect is also captured by the at least one camera;

[0016] The at least two fetches are asynchronous and executed randomly over a short period of time;

[0017] The central control and processing unit is capable of processing the acquisition and determining the presence or absence of counterfeit products after processing the acquisition using differentiated lighting.

[0018] In the context of this invention, "counterfeit" refers to the deceptive means described in the preamble of this application, and generally refers to any data that does not correspond to the actual face of the individual for whom a photographic portrait is to be obtained. Incidentally, the concept of counterfeit is described in standards ISO / IEC 30107-1 and -3.

[0019] Therefore, this invention enables a simple and rapid way to combat any deception attempt. In practice, typically, the total acquisition duration lasts less than one second, and these two acquisitions cannot be detected by the individual of interest or by any advanced electronic systems implemented by said individual, and are therefore unpredictable because the acquisitions occur randomly and in any order during this acquisition phase; that is, acquisitions under standard lighting can occur before or after acquisitions under differentiated lighting.

[0020] In the context of this invention, the concept of this random characteristic can be expressed in different ways.

[0021] Therefore, as mentioned above, the randomness can originate from the order in which the sequences were acquired (using standard lighting and differentiated lighting, respectively), and also from the duration of the pauses:

[0022] ● The interval between two consecutive acquisitions is typically close to tens of milliseconds;

[0023] ● and / or between the initialization of the acquisition process (i.e., the start of the cycle for acquiring the portrait of the individual of interest) and the valid start of the acquisition sequence.

[0024] Randomness can also stem from the duration of acquisition of the face of interest (using standard lighting and differential lighting, respectively), typically ranging from 1 ms to 30 ms.

[0025] This random characteristic can also originate from the variable frequency of the PWM, which is directly reflected in the activation of the LEDs that form part of the light source.

[0026] These different means designed to provide the randomness of the acquisition can be further combined to further enhance that randomness and thus significantly limit any risk of deception.

[0027] Furthermore, and in order to mislead individuals attempting such deception, the acquisition sequence may include emitting flashes of light from a light source in a manner unrelated to the at least two acquisitions (i.e., independent of the actual acquisitions). This makes the fraud analysis of the periodicity of the acquisitions even more complex.

[0028] According to the invention, two light sources are positioned laterally to the field of view of the camera. As a variation, the two light sources can be positioned differently, specifically above and below the field of view, to illuminate the face of interest from the top and bottom, or also according to a diagonal orientation, with the aim of emphasizing the prominence of the face by creating shadows within it.

[0029] According to the present invention, differentiated lighting corresponds to:

[0030] ● By weighted lateralization of facial lighting;

[0031] ● The aliasing effect is achieved through synchronization, and especially through the coordination and adaptation of camera parameters and PWM system parameters;

[0032] ● Partial illumination is provided by the at least two light sources in the infrared or even in wavelengths other than the visible spectrum; in this configuration, a second digital camera is provided with a sensor whose sensitivity is adapted to the wavelength or wavelength range under consideration.

[0033] ● The change in light intensity generated by the at least two light sources;

[0034] ● Alternatively, the lateralization of facial lighting and the 2D capture of an image with shadow variations according to the depth of the subject's face provide a depth effect and thus provide 3D capture (three-dimensional) of the face.

[0035] According to a first embodiment of the invention, the acquisition of an individual's face under differentiated lighting is achieved by weighting the illumination, and particularly by lateralizing the illumination. This lateralization is possible because the intensity of the two light sources can be modulated during the acquisition using a PWM system. In other words, the individual's face is illuminated at different intensities on either side of a midplane extending through the center between the eyes and the bridge of the nose. Processing of the acquired image makes it possible to distinguish all images that are not three-dimensional, such as portraits printed on flat or curved sheets, or images appearing on electronic displays (such as tablets or smartphones). In fact, due to the implementation of lateralized lighting, if a three-dimensional object is actually present, shadows caused by the protrusion of the face will appear at the level of the acquired image. According to this embodiment, it is also possible to change the aforementioned random characteristics by varying a variable ratio of the lateralization intensity (typically within a ratio from 1 to 5).

[0036] According to a second embodiment of the invention, the acquisition of an individual's face under differentiated lighting originates from the formation of an aliasing effect, which can be detected by the acquisition camera, and again enables the differentiation of images from tablet or smartphone types. More specifically, the parameters of the camera, and particularly its scanning speed, are synchronized with, and more particularly coordinated with, the PWM system acting on the light source (and therefore on the illumination frequency of the LED forming the light source).

[0037] In the presence of genuine individuals, the aliasing effect is clearly visible in the image captured by the camera, and across the entire surface area before illumination. However, in a configuration where the background of the compartment is effectively illuminated (see below), the aliasing effect is significantly attenuated against the background of the compartment if the background provides sufficiently significant light power. The background of the counterfeit (appearing on printed materials or electronic displays such as tablets or smartphones) and the background of the compartment exhibit different patterns, making them identifiable.

[0038] According to a third embodiment of the invention, the acquisition of an individual's face under differentiated illumination is achieved, for example, through partial illumination in the infrared. For this purpose, at least one of the LEDs in at least one of the two light sources emits infrared light. This partial illumination in the infrared makes it possible to detect the presence of a real face, and more specifically, the presence of skin on said face. In fact, the reflectivity of the support and, for example, the skin, varies with wavelength, thus enabling the differentiation of skin.

[0039] According to a fourth embodiment of the invention, the acquisition of an individual's face under differentiated lighting is based on variations in light intensity generated by the at least two light sources. Typically, if the light intensity of the light sources doubles, this doubling occurs on the subject's face. However, due to the light intensity generated by the display, this doubling does not occur on the electronic display held by the subject. The processing unit enables the detection of the presence or absence of this proportionality and the corresponding identification of counterfeits.

[0040] According to another embodiment of the invention, the acquisition of an individual's face under differentiated lighting is achieved by a simple lateralization of the lighting, which in effect generates a shadow effect on one of the facial features, typically defined by a vertical line extending along the bridge of the nose. This lateralization is made possible by the aforementioned two light sources, which are laterally positioned relative to the field of view of the acquiring camera and, for example, symmetrical with respect to the field of view of the acquiring camera. The camera can be of a standard 2D type and is therefore able to acquire images along the depth axis by means of the shadow effect, rather than in a restrictive manner along a plane perpendicular to the optical axis of the camera. Then, by means of suitable software, pixel-by-pixel calculations are performed on the two images acquired consecutively by the camera, for each primary color (red, green, blue), and thus a 3D mapping is achieved that enables the differentiation between forgeries and genuine faces.

[0041] According to another aspect of the invention, the device of the type of interest may include a background of a wall facing the compartment, integrating an acquisition camera, the background also having an illumination source managed by a central unit. In this configuration, by managing the frequency of the background illumination and the camera scanning frequency, the acquisition of an individual's face under differential illumination achieves the formation of an aliasing effect at the horizontal level of the background. In fact, in this configuration, the background illumination source is positioned behind the subject, such that the aliasing effect appears only on the background. Therefore, if the subject has an electronic display, the captured background does not exhibit aliasing. Thus, individuals actually present within the device can be distinguished based on the screen of a tablet or smartphone.

[0042] Regardless of the differentiated lighting patterns that are achieved and may accumulate, the images thus acquired by the camera are processed by a central management and processing unit that uses artificial intelligence to distinguish between an actual individual in front of the camera and a 2D image or reconstruction, or even from the placement of a mask on the face of the relevant individual.

[0043] In other words, the present invention generally includes modulating the illumination of the face of an individual whose portrait is to be acquired in order to combat deception (especially inherently for those wearing masks), expose the screen of a smartphone or tablet or printed sheet, or even combat “distortion”.

[0044] According to a variation of the invention, it is conceivable to add two additional captures or acquisitions under the same lighting conditions. Thus, and by means of a suitable algorithm managed by a central control and processing unit, it is possible to determine the stability of individuals along the entire integrated sequence, thereby making the presentation of any artifacts impossible during said sequence.

[0045] The present invention also relates to an apparatus for acquiring a photographic portrait of an individual to implement the aforementioned method. Such an apparatus is formed by a compartment defining an interior space in which the individual may be positioned, and the apparatus includes:

[0046] ■ A first wall, defining a portion of the interior space of the compartment, at least a portion of which is transparent;

[0047] ■ At least one digital camera is arranged behind a transparent or translucent portion of the wall and includes a field of view oriented toward the interior of the compartment, with its focal plane located within the interior space of the compartment;

[0048] ■ At least two light sources illuminate the interior of the compartment and are arranged such that they provide side illumination of the individual's face and are located outside the field of view of the camera.

[0049] According to an embodiment, these illumination sources are formed by light-emitting diodes (LEDs) managed by PWM, and the at least one camera is digital. The management of the illumination of the LEDs and the acquisition by the camera is managed by a central control unit, which also ensures the processing of images acquired by the camera.

[0050] According to a variation of the invention, the aforementioned device further includes a background opposite the wall in which the camera is integrated, the background being provided with an illumination source also managed by the PWM system.

[0051] According to a variation of the invention, the compartment can be replaced by a support frame (i.e., a lighter structure).

[0052] According to another variation of the invention, the entire process of acquiring an individual's portrait is typically supervised by a surveillance agency. In other words, when secure registration is desired, as may be required in many applications, an operator outside the device must be able to monitor the different steps of the acquisition. For this purpose, the device may include an additional camera designed to remotely transmit the contents of the compartment in real time. As a variation, at least one of the walls defining the interior space of the compartment may be provided with variable light-blocking properties. Attached Figure Description

[0053] The invention and its advantages will become more apparent from the following non-limiting embodiments, taken in conjunction with the accompanying drawings.

[0054] Figure 1 It is a simplified representation of a cross section of an embodiment of a photographic portrait acquisition device according to the present invention.

[0055] Figure 2 yes Figure 1 A simplified representation of the front view of one of the walls of the device.

[0056] Figure 3 It is an alternative embodiment of lighting, similar to Figure 2 A simplified view.

[0057] Figure 4 This is a simplified top view showing the first differentiated lighting pattern in the case of lateral lighting.

[0058] Figure 5 An image of the face is schematically shown after the lighting has been lateralized.

[0059] Figure 6 Showing under normal lighting Figure 5 The image.

[0060] Figure 7 It is a simplified view of the face image obtained under the aliasing of the second differential lighting mode.

[0061] Figure 8 Is with Figure 7 The image is a simplified view of the smartphone screen obtained under the comparison of the overlay.

[0062] Figure 9 This illustrates another differentiated lighting pattern similar to that shown in the case of partial illumination under infrared conditions. Figure 4 A simplified top view.

[0063] Figure 10 An image of a face acquired under the aliasing of a background in a compartment, according to another differentiated lighting pattern, is schematically shown.

[0064] Figure 11 It shows the changes in the parameters that generate aliasing, similar to Figure 10 The view.

[0065] Figure 12 In the context of Figure 10 Under the same conditions, obtain a simplified view of the smartphone screen.

[0066] Figure 13 In the context of Figure 10 Obtain a simplified view of the printed sheet under the same conditions.

[0067] Figure 14 This is a simplified view showing an image of the face acquired under the aliasing of background and side lighting in the compartment (i.e., a mixture of two previously shown differentiated lighting patterns).

[0068] Figure 15 It is a simplified view showing the acquired face image in the previous mode, with the camera saturated by background lighting.

[0069] Figure 16 It shows something similar to Figure 15 A simplified view of the printed sheet image obtained under the same conditions.

[0070] Figure 17 It is similar to Figure 16 The text appears to be a mix of Chinese characters and symbols, possibly representing a corrupted or incomplete document. A direct translation wouldn't be meaningful without further context or clarification. Figure 15 Images of smartphones acquired under the same conditions.

[0071] Figure 18 It is a simplified view showing a user's face printed on a curved sheet in depth mode and translated through 3D mapping.

[0072] Figure 19 It is similar to Figure 18 It obtains a view of the user's real face in depth mode. Detailed Implementation

[0073] Figure 1 An example of an embodiment of a photographic portrait acquisition device 10 according to the present invention is shown. In addition to the top panel 24, the device 10 also includes a compartment 12 defined by four walls 14, 16, 18 and 20; one of the walls is only partial to allow an individual to be introduced into the overall volume defined by the wall.

[0074] The first wall 14 includes at least a transparent portion so that the digital camera 22 positioned behind the wall can have its field of view 23 and its focal plane 26 within the compartment.

[0075] Seat 25 may be present in the compartment so that user 30 can sit in front of and close to the focal plane 26 of digital camera 22. More precisely, the seat is positioned to include the focal plane of the camera. Adjustment devices allow the user to adjust the position of the seat or camera to position his / her face 31 in a reference position 32. The reference position is centered in the plane defined by the intersection of the field of view 23 of digital camera 22 and the focal plane 26.

[0076] Device 10 includes components that enable a user to position his / her face in a reference position, whether he / she is standing or sitting in seat 25 (if present). These components may be visual indicators on wall 14 located in front of digital camera 22, such as markers to position his / her eyes at the correct height in the camera's field of view.

[0077] The device includes a light source 35 positioned between the focal plane 26 of the digital camera 22 and the first wall 14 of the compartment. According to this example, the light source 35 is attached to and embedded therein by means of suitable components on the top plate 24 of the compartment.

[0078] like Figure 2 As shown, the light source 35 includes a first light-emitting element and a second light-emitting element, effectively defining two light sources. More precisely, the first light-emitting element 36 is arranged to illuminate the second wall 16 that defines the compartment and is adjacent to the first wall 14, and the second light-emitting element 37 is positioned to illuminate the third wall 18 of the compartment facing the second wall 16. The second and third walls are thus positioned relative to each other and intersect the first wall 14, orthogonally or with an obtuse angle defining the space between them. In this latter configuration, the cross-section of the compartment is no longer rectangular but trapezoidal.

[0079] according to Figures 1 to 3 In the embodiment shown, the first and second light-emitting elements are each formed by the alignment of LEDs 38, which are held and oriented in front of the side walls 16, 18 of the compartment, for example via corresponding first spacers 39 and second spacers 40. However, any other arrangement is conceivable. In other words, the longitudinal arrangement of the LEDs is substantially parallel to the second wall 16 and / or the third wall 18. The LEDs emit warm white light with a color rendering index of 85, a temperature of approximately 4000 Kelvin, an intensity of at least 550 lumens (per light-emitting element), and a diffusion angle between 30° and 40°. These different LEDs are powered by appropriate current sources, and their power supplies are periodically switched by a PWM system, for which the pulse frequency ranges from a few hertz to several kilohertz and the duty cycle varies from 0 to 100% depending on the desired effect. The PWM system itself is managed via an electronic interface board by a central control and processing unit (not shown) via a USB interface.

[0080] More specifically, according to a preferred embodiment of the invention, the angles of the shims 39 and 40 are selected such that the light-emitting elements 36 and 37 illuminate the areas of the second wall 16 and the third wall 18, which are located at the same height in the compartment as the surface defined by the intersection of the field of view 23 of the digital camera 22 and the focal plane 26. The tilt angles generated by the shims for each light-emitting element 36, 37 may not be exactly the same, and the optimization of these angles typically depends on the distance between the user (reference position 32) and the walls 16 and 18, the distance between the user and the height of the light source 35, and the distance between the axes of the two light-emitting elements 36 and 37.

[0081] Thus, the light emitted by the light source 35 is reflected in regions 41 and 42 of the walls 16 and 18, forming a first diffuse luminescent surface on the second wall 16 and a second diffuse luminescent surface on the second wall 18. These diffuse luminescent surfaces are positioned in front of and to the sides of the user's face 31. The light generated is also reflected on the first wall 14, and these different walls are coated with appropriate components capable of reflecting light.

[0082] Advantageously, the light-emitting surfaces 41, 42, and 14 illuminate all parts of the user's face according to a large number of incident directions. As a result, the face is uniformly illuminated.

[0083] In addition to the technical characteristics (light intensity) of these two light-emitting elements, the device of the present invention can also eliminate the need for any facial light source due to their presence, their position, and their operating mode, and especially due to the reflection of light emitted by them on the diffuse surfaces 41, 42, and 14, provided that the inner surface of wall 14 is coated with a reflective layer. Therefore, a portion of the light emitted by light sources 36 and 37 is reflected on wall 14 and used as facial (or frontal) illumination for the user's face.

[0084] However, the following facial light source can be achieved. It is already... Figures 1 to 3 Figure 45 is shown in the figure. The facial light source can be formed by LED 46, which emits warm white light characterized by a color rendering index greater than 85, a temperature of approximately 4,000 Kelvin, a natural intensity of greater than 550 lumens, and a diffusion angle of 120°.

[0085] In practice, the presence of two side light sources 36 and 37 reduces the intensity of light emitted by the facial light source 45 (when the facial light source is present), or the absence of the facial light source can be overcome by changing the illumination intensity of the side light sources 36 and 37. Consequently, the facial light source 45 causes less or even no glare to the user, thus providing better user comfort.

[0086] According to the invention, a fourth wall 20 encloses the compartment and forms a background opposite to the first wall 14. The fourth wall 20 may be at least partially transparent, allowing another emitting device 50 disposed behind the wall to illuminate the interior of the compartment. The emitting device 50 may comprise an array network of LEDs 51 of different colors (e.g., blue, red, and green), enabling the background of the compartment to be illuminated according to multiple colors. The LEDs may be similar to those described above. The array of LEDs 51 is also managed by the aforementioned PWM system.

[0087] The surface of the fourth wall 20 may also be covered with an anti-reflective material (or pad) and a dark layer, such as black and translucent. This surface of the fourth wall can then absorb a portion of the light originating from inside the compartment to achieve optimal color saturation and uniformity of the background of the compartment through the fourth emitting device.

[0088] Furthermore, at least one of the two side walls 16, 18 of the compartment must allow the user to be drawn into its defined volume. For this purpose, one of these two walls 16, 18 is either slidable or replaced by a curtain or any light-blocking support. In the latter case, and taking into account the difference in nature between such a curtain and the side wall facing the curtain, and in order to have adequate lighting, the light sources 36, 37 are varied by specifically differentiating their intensities via a PWM system.

[0089] Furthermore, in the context of the implementation of such a compartment for secure registration (meaning that the acquisition and processing of the relevant user's face is supervised by a third party (e.g., by an operator designated for this purpose), the side walls 16, 18 and the background wall 20 may be provided with variable light-blocking properties.

[0090] Typically, all or part of the wall may be made of a transparent material (e.g., glass) and coated with a liquid crystal film. This film has electro-optic properties, allowing it to switch from a light-blocking state to a transparent state when an electric field is applied. In this case, a central control and processing unit synchronizes the electrical pulses controlling the transition between the light-blocking and transparent states with at least two acquisitions of the face to ensure the desired illumination. The actual acquisition phase is performed in the light-blocking state.

[0091] It should be understood that such a membrane can be replaced by any system that ensures the same function, and is, for example, a Venetian or Californian blind type, whose orientation is controlled by a motor and also instructed by a control and processing unit.

[0092] According to a variation of the previous embodiment, and still for the purpose of security registration, a surveillance camera is positioned in the compartment and, for example, on the ceiling, which captures images of the interior of the compartment in real time, thereby enabling the operator or monitoring agency to remotely control these operations.

[0093] A method for obtaining biometric photographic portraits of individuals using the previously described equipment will now be described.

[0094] According to the essential features of the invention, once the relevant individual is in place in the compartment, at least two acquisitions of the individual's face are performed:

[0095] ●Acquisition performed under standard lighting, wherein two light sources 36, 37 are activated simultaneously to uniformly illuminate the face so as to allow biometric acquisition of the face by digital camera 22 for subsequent use, for example, for identity verification or access control support;

[0096] ●And the acquisition performed under differentiated lighting, namely, the lighting effect generated on the surface by two light sources 36, 37, is also acquired by the digital camera 22.

[0097] These two fetches are not executed synchronously, but they are executed randomly over a short period of time (typically less than 1 second, and preferably less than 300 milliseconds).

[0098] The at least two acquisitions are processed by a control and processing unit integrated in the compartment, and the control and processing unit is able to determine whether counterfeit products exist after processing the acquisitions performed according to the differentiated lighting.

[0099] The different differentiated lighting patterns will now be described.

[0100] According to Figures 4 to 6 The first mode of the invention shown here, wherein such differentiated lighting is formed by weighting, and more precisely by lateralizing, the lighting. This lateralization is made possible by two light sources 36, 37 (the intensity of which can be modulated during acquisition by means of a PWM system). In other words, two portions of an individual's face, substantially located on either side of the midplane extending through the center between the eyes and the bridge of the nose, are differentiated by different intensities (as shown in...). Figure 4 (The light source is illuminated by arrows of unequal lengths starting from light source 37).

[0101] Therefore, obtain Figure 5 The image shown effectively illustrates the lighting difference between the left and right sides of the face. This can still be achieved by using a PWM system to change the intensity difference between the two light sources 36 and 37.

[0102] The processing of the acquired images thus enables the differentiation of all images that are not three-dimensional, including, for example, portraits printed on flat or curved sheets, or images appearing on electronic displays (such as tablets or smartphones). In fact, due to the implementation of lateral lighting, if effectively present in the presence of a three-dimensional object, a difference in brightness will appear at the level of the acquired image due to shadows caused by the protrusion of the face. Figure 5 ).

[0103] Typically, with the aid of appropriate algorithms, the central processing unit is able to decompose an image of the facial texture (which may exist, for example, on a printed photograph) and a depth image of the face on the other hand. Thus, with the aid of appropriate algorithms for this depth image of the face, analysis through artificial intelligence enables the detection of flat or curved supports of the type of printed sheet, tablet computer, or smartphone screen appearing in front of the camera 22.

[0104] According to Figure 7 and 8 The second mode of the invention shown achieves this differentiated lighting by creating an aliasing effect that can be detected by the digital camera 22.

[0105] In this case, the parameters of the digital camera 22, particularly its scanning speed, are synchronized and more precisely coordinated with the PWM system, which acts on the light sources 36 and 37, and thus on the illumination frequency of the LED 38 forming the light source. For this purpose, the camera 22 implements a CMOS sensor, which is better referred to as a "rolling shutter," and is shown, for example, at the internet link https: / / en.wikipedia.org / wiki / Rolling_shutter.

[0106] This aliasing effect appears as stripes on the acquired image, particularly on the individual's face. Figure 7 The width of these stripes depends on the pulse duration of LED 38 (known to be related to the frequency and duty cycle of the PWM system, i.e., the intensity variation of LED 38) and the line readout duration of digital camera 22. The nature of these stripes can vary in intensity and width over certain areas of the image. These areas particularly relate to light-emitting surfaces, possibly in front of camera 22. Therefore, typically, electronic displays of smartphones or tablets, or even illuminated surfaces (such as surface 20 when surface 20 is equipped with a light source 50), affect the image captured by the camera. Thus, in Figure 8 As can be observed, the area limited to the smartphone screen (i.e., the luminous area) does not exhibit aliasing. Therefore, the central processing unit is able to distinguish the presence of a real individual or the presence of an imposter caused by the image being presented in front of the camera 22, and generally distinguish the light source.

[0107] According to Figure 9 The third embodiment of the invention shown here, wherein the differentiated illumination is formed by partial illumination of the individual's face in a wavelength spectrum other than visible light, and for example in the infrared. For this purpose, at least one of the LEDs 38 of at least one or even two light sources 36, 37 emits infrared light. Figure 8 (The two top arrows in the image).

[0108] This partial infrared illumination enables the detection of a real face, and more specifically, the presence of skin on that face. In fact, the reflectivity of the support, and skin, varies with wavelength, thus making it possible to distinguish skin at the processing unit level using artificial intelligence relative to more neutral supports (such as sheets of paper or images from smartphones or tablets). Therefore, typically, photographic paper representing a face, perhaps for deceptive purposes, may have a significantly higher reflectivity than a real face.

[0109] According to a fourth mode (not shown) of the invention, this differentiated illumination is achieved through a variation in the light intensity generated by two light sources 36 and 37. Typically, if the light intensity of the light sources is doubled, this doubling appears on the subject's face and is detected by the central processing unit. However, due to the light intensity generated by the electronic display of a smartphone or tablet computer held by the subject, this doubling of intensity will not appear on the display. The processing unit enables the detection of the presence or absence of this proportionality and the corresponding identification of counterfeits.

[0110] According to the combination Figures 10 to 13 The fifth mode of the invention shown here achieves differentiated lighting by forming an aliasing at the level of the background 20 of the compartment when the compartment is equipped with a lighting source 50 of the type previously discussed.

[0111] Therefore, by changing the scanning frequency of the digital camera 22 (again, a "rolling shutter" type) and the illumination frequency of the LED 51 of the light source 50 of the background 20, the aliasing effect can only be observed on the background of the acquired image. Figure 10 and Figure 11 Therefore, if the subject remains on the printed paper sheet ( Figure 13 If an electronic display (compartment 12) is used, the background captured locally around the subject will not appear aliased. Therefore, in Figure 12 and 13 In this study, it can be observed that there is no overlay on the area behind the face appearing on the screen or paper. This demonstrates that the central processing unit is able to distinguish the actual individual within the device from the screen of a tablet or smartphone.

[0112] According to the present invention, these different differentiated lighting patterns can be implemented individually or in combination. This combination optimizes the central processing unit's ability to distinguish counterfeits, and thus still optimizes the ability to more effectively resist any deception attempts.

[0113] For example, already about Figures 14 to 17 A combined implementation of the previously described second and fifth embodiments is described, namely the two aliasing modes described.

[0114] Therefore, in Figure 14 In this study, double aliasing can be observed, i.e., different jump widths between the individual and the rest of the acquired data.

[0115] exist Figure 15 In this process, the intensity of the light source 50 on the background wall 20 is changed to saturate the digital camera 22. Therefore, only... Figure 15Aliasing was observed on individual samples, with the remainder of the acquired image appearing white. In the same configuration, i.e., with the digital camera 22 saturated by the intensity of the light source 50, the acquisition of the printed paper sheet was shown... Figure 16 It can be observed that the entire paper sheet (including the area of ​​the sheet outside the printed portrait) exhibits overlapping. Similarly, and in the same configuration, in Figure 17 It can be observed that the smartphone is only covered by aliasing on the periphery of the actual image, which may also generate saturation of the digital camera 22.

[0116] This combination of differentiated lighting patterns (distinguished by the central processing unit) optimizes anti-spoofing.

[0117] exist Figure 18 and 19 Another mode of the invention is shown, which is based on the principle of lateralization and directionalization of lighting on the user's face in the context of differentiated lighting.

[0118] In these figures, the first row corresponds to the original acquisition under differential lighting, the second row corresponds to the extraction of the face, and the third row is the extraction of the depth map for different color planes.

[0119] Therefore, in Figure 18 In the image, there is a deception attempt by displaying a cylindrically curved printed sheet of paper in front of the camera (typically a 2D camera) showing a face (first line). The second line schematically illustrates an image captured by the camera by altering the illumination emitted from a side light source, thereby assuming the creation of shadow areas relative to a false line extending through the center between the eyes and the bridge of the nose. The captured image is then processed by an algorithm for determining the illumination difference for each pixel acquired in the acquisition. This processing produces the result shown in the third line, where the system detects the deception attempt due to the use of the curved printed sheet of paper.

[0120] This Figure 18 and Figure 19 In comparison, Figure 19 In the image, the real user is positioned in front of the camera. The image capture algorithm enables the display of a real individual (in the third row), thus avoiding the risk of deception.

[0121] The previously different differentiated lighting patterns can be further linked to enhancements of acquired random characteristics, such as those specified in the context of the above statements of this invention.

[0122] The advantages of this invention in combating fraud are understandable, as fraud is a highly critical topic for security services (whether government or private).

Claims

1. A method for acquiring a photographic portrait of an individual, wherein the individual is within an interior space defined by a compartment, the method comprising: Biometric images of the individual's face are acquired using at least one digital camera (22). Simultaneously with this acquisition, the face of the individual is illuminated by at least two separate light sources (36, 37) that are not centered on the face, wherein the at least two light sources (36, 37) are formed by light-emitting diodes (38) managed by a PWM system; The PWM system and the operation of the at least one digital camera (22) are managed by a central control and processing unit. In the method, at least two acquisitions of the face are performed respectively: Using standard lighting, wherein the at least two light sources (36, 37) are simultaneously activated to uniformly illuminate the face, thereby allowing the at least one digital camera (22) to acquire biometric data of the face; Differential lighting is used, wherein the at least two light sources (36, 37) generate a lighting effect on the face, and the lighting effect is also captured by the at least one digital camera; The at least two acquisitions are asynchronous and executed randomly within a time span of less than 1 second; and The central control and processing unit processes the at least two acquisitions and determines whether counterfeit products exist after processing the acquisition using differentiated lighting.

2. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The at least two light sources are positioned laterally toward the field of view (23) of the at least one digital camera (22) or above and below the field of view of the at least one digital camera (22), or are also positioned to illuminate the face of interest according to a diagonal orientation, thereby emphasizing the prominence of the face by making shadows appear in the face.

3. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The differentiated lighting, individually or in combination, originates from the following: The aliasing effect is achieved through synchronization; At least one of the at least two light sources (36, 37) provides partial illumination in a wavelength spectrum outside the visible spectrum; The change in light intensity generated by the at least two light sources (36, 37); Lateralization of facial illumination achieved by the at least two light sources (36, 37).

4. The method for obtaining a photographic portrait of an individual according to claim 3, wherein, The aliasing effect is achieved by coordinating and adapting the parameters of the at least one digital camera (22) and the parameters of the PWM system.

5. The method for obtaining a photographic portrait of an individual according to claim 4, wherein, The acquisition of the individual's face under differentiated lighting originates from the formation of an aliasing effect detectable by the at least one digital camera (22), the formation of which originates from the synchronization of the parameters of the at least one digital camera (22) and the parameters of the PWM system.

6. The method for obtaining a photographic portrait of an individual according to claim 3, wherein, Acquisition of the individual's face under differentiated illumination achieves partial infrared illumination, wherein at least one LED forming part of each of the at least two light sources (36, 37) emits infrared light.

7. The method for obtaining a photographic portrait of an individual according to claim 3, wherein, Acquiring the individual's face under differentiated lighting is achieved by lateralizing the lighting and generating a shadow area at the level of the face, wherein the lateralization of the lighting originates from the management of the at least two light sources (36, 37) by means of the PWM system, and the shadow area is located on one side or the other side extending through the midplane between the eyes and the bridge of the nose, wherein the at least one digital camera is a 2D camera, the image acquired after the lateralization is captured by the 2D camera, and thus it is possible to acquire an image along the depth axis through the shadow effect, and then the image is processed pixel by pixel to generate a 3D map that can distinguish between fake and real faces.

8. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, An additional lighting source (50) is also formed of LEDs and faces the at least one digital camera (22) and is positioned outside the focal plane (26) of the at least one digital camera (22). The additional lighting source (50) is also managed by the central control and processing unit, and wherein the acquisition of the individual's face under differentiated lighting originates from the formation of aliasing at the background level of the image acquired by the at least one digital camera (22). The formation of this aliasing effect originates from synchronization.

9. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The randomness of the acquisitions using standard lighting and differentiated lighting is related to the order of the acquisitions.

10. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The randomness of acquisition using standard lighting and differentiated lighting is correlated with the duration of the pauses, which are respectively: A pause between two consecutive acquisitions, the duration of which is close to tens of milliseconds; and / or The initialization of the acquisition process is the pause between the start of the acquisition cycle for the portrait of the individual of interest and the effective start of the acquisition sequence.

11. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The randomness of acquisition using standard illumination and differentiated illumination was correlated with the duration of acquisition of the face of interest using standard illumination and differentiated illumination, respectively, where the duration ranged from 1 ms to 30 ms.

12. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The randomness of the acquisition using standard lighting and differentiated lighting is associated with the variable frequency of the PWM system, which is directly reflected in the activation of the LEDs that form part of the at least two light sources.

13. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The randomness of the acquisition using standard lighting and differentiated lighting is associated with the at least two light sources (36, 37) emitting flashes in a manner unrelated to the at least two acquisitions.

14. The method for obtaining a photographic portrait of an individual according to claim 1, wherein, The at least two fetches are performed randomly within a time period of less than 300 milliseconds.

15. The method for obtaining a photographic portrait of an individual according to claim 3, wherein, The partial illumination refers to partial illumination in the infrared spectrum.

16. The method for obtaining a photographic portrait of an individual according to claim 3, wherein, Acquisition of the individual's face under differentiated lighting is achieved by lateralizing the lighting and generating lighting of different intensities at the level of the face on two parts of the individual's face located on both sides of the mid-plane extending through the center between the eyes and the bridge of the nose, wherein the lateralization of the lighting originates from managing the at least two light sources (36, 37) by means of the PWM system.

17. The method for obtaining a photographic portrait of an individual according to claim 5, wherein, The aliasing effect is formed by the coordination and adaptation of the operating parameters of the at least one digital camera (22) with the PWM system, wherein the PWM system acts on the at least two light sources (36, 37) and thus on the illumination frequency of the LEDs (38) forming the at least two light sources.

18. The method for obtaining a photographic portrait of an individual according to claim 8, wherein, The aliasing effect is formed by the coordination and adaptation of the parameters of the at least one digital camera (22) with the PWM system, wherein the PWM system acts on the LED (51) of the additional lighting source (50), and thus on the lighting frequency of the LED (51) forming the additional lighting source.

19. An apparatus for acquiring a photographic portrait of an individual, the apparatus comprising a compartment (10) defining an interior space (12) in which the individual is positioned, the apparatus comprising: A first wall (14) defines a portion of the interior space of the compartment, at least a portion of which is transparent; A camera (22) is arranged behind the transparent portion of the first wall (14) and includes a field of view (23) oriented toward the interior of the compartment, and the focal plane (26) of the camera is located in the compartment; At least two light sources (36, 37) illuminate the interior of the compartment and are arranged such that the at least two light sources illuminate the face (31) of the individual laterally, and the at least two light sources are located outside the field of view (23) of the camera (22). The at least two illumination sources (36, 37) are formed by light-emitting diodes (LEDs) (38) managed by a PWM system. The camera (22) is digital, and the management of the illumination of the LEDs (38) and the management of the acquisition by the camera (22) are managed by a central control unit. The central control unit also ensures the processing of images acquired by the camera. The at least two illumination sources are activated synchronously to allow the camera to acquire biometric features of the individual's face. The at least two illumination sources generate differentiated illumination to allow the camera to acquire additional features of the individual's face. The acquisition is not performed synchronously and is performed randomly over a time period of less than 1 second.

20. The apparatus for acquiring a photographic portrait of an individual according to claim 19, wherein, Each of the at least two illumination sources (36, 37) is formed by the alignment of LEDs (38), the at least two illumination sources (36, 37) are positioned at the level of the top plate (24) of the device, between the first wall (14) and the focal plane (26) of the camera (22), and the at least two illumination sources (36, 37) illuminate the side walls (16, 18) of the device that intersect with the first wall (14).

21. The apparatus for acquiring a photographic portrait of an individual according to claim 20, wherein, At least one of the sidewalls (16, 18) is formed by a retractable curtain or a light-blocking support.

22. The apparatus for acquiring a photographic portrait of an individual according to claim 20, further comprising a background wall (20) opposite to and positioned outside the focal plane (26) of the camera (22) integrated therein, the background wall (20) being provided with an additional illumination source (50) formed by LEDs (51) and managed by the PWM system, the PWM system also ensuring management of the at least two illumination sources (36, 37).

23. The apparatus for acquiring a photographic portrait of an individual according to claim 22, wherein, At least one of the sidewalls (16, 18) and the background wall (20) is transparent and has variable light-blocking properties, and the state change between the transparent state and the light-blocking state is managed by the central control unit.

24. The apparatus for acquiring a photographic portrait of an individual according to claim 23, wherein, At least one wall of the compartment is coated with a liquid crystal film, which switches from a light-blocking state to a transparent state when an electrical pulse synchronized by the central control unit is applied.

25. The apparatus for acquiring a photographic portrait of an individual according to claim 19, wherein, The compartment is integrated with an additional camera for remotely transmitting the contents of the compartment in real time.

26. The apparatus for acquiring a photographic portrait of an individual according to claim 19, wherein, The acquisition is performed randomly within a time period of less than 300 milliseconds.

Images