Facial recognition system for the face of a motor vehicle driver

Abstract

Title of the invention: Facial recognition system for the face of a motor vehicle driver. Facial recognition system (10) for the face (5) of a motor vehicle driver (100), the facial recognition system (10) comprising a light source (12) configured to produce a light beam (F) in the infrared range and an optical device (14) comprising a metal lens (16), the optical device (14) being disposed in the path of the light beam (F) produced by the light source (12), the metal lens (16) being configured to diffract the light beam (F) so as to project at least one pattern into a field of vision and in particular onto the face (5) of a motor vehicle driver (100). Figure for the abstract: Fig. 1

Description

Title of the invention: Facial recognition system for the face of a motor vehicle driver

[0001] The present invention relates to a facial recognition system for identifying the face of a motor vehicle driver. The invention also relates to a motor vehicle equipped with such a facial recognition system. The invention further relates to a computer program implementing the aforementioned method. Finally, the invention relates to a recording medium on which such a program is recorded.

[0002] The automotive industry has been developing the use of facial recognition to activate various vehicle functions for some time. Some functions require higher or lower levels of security. For example, facial recognition processing used as a vehicle anti-theft function will require a higher level of security than facial recognition processing used for personalized driver welcome.

[0003] Sensors commonly used in this type of application include, for example, RGB cameras and / or infrared cameras. These sensors allow for capturing images of a person's face in the driver's seat of a motor vehicle. Infrared camera technology, for instance, allows for the analysis of reflections present in the images, while authentication using RGB camera images relies on the detection of eye or facial movements, blood pulse, skin texture, or facial geometry. RGB cameras and infrared cameras are therefore sensors that each have their own specific operating mechanisms and can be used in a complementary manner to perform facial recognition processing.

[0004] These authentication solutions therefore require specific installations for each of the sensors. Furthermore, the field of view of the components in the solutions described above is relatively limited, which restricts their scope of application. In addition, these solutions are generally bulky, heavy, and poorly suited for use in an environment such as a motor vehicle's interior.

[0005] The object of the invention is to provide a facial recognition system that overcomes the above drawbacks and can be easily integrated into a motor vehicle. In particular, the invention allows for increased compactness and lightness while offering a wider field of view than existing solutions, without any loss of image quality.

[0006] To this end, the invention relates to a facial recognition system for the face of a motor vehicle driver, the facial recognition system comprising a light source configured to produce a light beam in the infrared range and an optical device comprising a metal lens, the optical device being disposed on the path of the light beam produced by the light source, the metal lens being configured to diffract the light beam so as to project at least one pattern in a field of vision and in particular on the face of a motor vehicle driver.

[0007] According to one embodiment, the facial recognition system further comprises: - at least one light receiving device including at least one analysis unit configured to perform facial recognition from the light reflected by the driver's face; - at least one memory unit configured to store at least one personalized driver profile; - at least one processing unit configured to compare the results of facial recognition performed by the light-receiving device with at least one personalized driver profile stored in the memory unit, - at least one executive unit configured to perform an action on at least one vehicle parameter based on a result from the comparison performed by the processing unit.

[0008] According to one embodiment, the pattern intended to be projected onto the face of a motor vehicle driver by the optical device comprises a set of luminous points having a uniform intensity.

[0009] According to one embodiment, the meta-lens comprises at least one nanostructure comprising a transparent substrate with an arrangement of nanopillars on its surface and the nanopillars have a size less than or equal to the wavelength of the light beam to diffract the light beam produced by at least one light source and passing through the meta-lens.

[0010] According to one embodiment, the nanopillars have a cylindrical shape and the diameter of the nanopillars is between 100nm and 500nm.

[0011] According to one embodiment, the nanopillars comprise a material with an optical index greater than 2, and more particularly greater than 3 or more particularly equal to 3.47.

[0012] The invention also relates to a method of operating a facial recognition system for the face of a motor vehicle driver, the method comprising the following steps: - emission of a light beam in the infrared range from at least a light source through at least one optical device comprising a metal lens and directed towards the face of a motor vehicle driver; - diffraction of said light beam by the metal lens of the optical device so as to project at least one pattern onto the face of the driver; - reception of light reflected by the driver's face in the light receiving device; - execution of facial recognition from said light by analysis unit 9 - comparison of the result of facial recognition with at least one personalized driver profile previously recorded in the memory unit by the processing unit; - implementation of an action on at least one vehicle parameter based on a result obtained from the comparison by at least one executive unit.

[0013] The invention also relates to a computer program product comprising program code instructions recorded on a computer-readable medium to implement the steps of the process described above when said program is running on a computer.

[0014] The invention also relates to a computer-readable data storage medium on which a computer program is stored, comprising program code instructions for implementing the method described above. (See figures)

[0015] These objects, features and advantages of the present invention will be described in detail in the following description of various particular embodiments, given by way of non-limiting example, in relation to the accompanying figures, among which:

[0016] Fig. 1 schematically illustrates the interior of a motor vehicle passenger compartment equipped with a facial recognition system according to the invention. Figure [Fig. 2] schematically illustrates a set of light points projected onto a driver's face with a state-of-the-art facial recognition system.

[0017] Fig. 3 schematically illustrates a set of light points projected onto the face of a driver with the facial recognition system according to the invention.

[0018] Fig. 4 is a schematic view of a nanostructure of a meta-lens.

[0019] Fig. 5 is another schematic view of a detail of a nanostructure of a meta-lens.

[0020] Figure [Fig. 6] schematically illustrates a method of implementing facial recognition from a photograph. Detailed description

[0021] An example of a motor vehicle 100 equipped with an embodiment of a facial recognition system is described below with reference to [Fig. 1]. To facilitate the representation of the different elements on [Fig. 1], the elements shown are not all at the same scale.

[0022] The motor vehicle 100 can be a vehicle of any type, for example a passenger vehicle or a utility vehicle, or a public transport vehicle such as a bus for example, or even a truck.

[0023] The motor vehicle 100 includes a passenger compartment 1 defining the interior space of the motor vehicle 100 and within which is arranged the facial recognition system 10 for the face 5 of a motor vehicle driver. This facial recognition system 10 is more particularly arranged at the front of the passenger compartment 1, preferably in front of the driver's seat so that the facial recognition system 10 can be used on the person who is sitting in that driver's seat.

[0024] The facial recognition system 10 includes at least one light source 12, at least one optical device 14 including a metal lens 16, at least one light receiving device 18 including at least one analysis unit 19, at least one memory unit 20, at least one processing unit 22 and at least one executive unit 24.

[0025] At least one light source 12 is configured to produce a monochrome light beam F. In particular, at least one light source 12 is configured to produce a light beam F in the infrared range. This type of light is invisible to the human eye, which has the advantage of not distracting the person being identified when this light source 12 illuminates the head, and more particularly the face 5, of said person. The light source 12 may emit pulsed light. Alternatively, the light source 12 may emit a continuous beam. The wavelength of the infrared light emitted by the light source 12 is greater than or equal to 850 nm; more particularly, it may be equal to 940 nm, or it may be equal to 1550 nm.

[0026] The optical device 14 of the facial recognition system 10 is positioned in the path of the infrared light beam F produced by the light source 12. At least one optical device 14 comprises a metal lens 16 through which the infrared light beam F passes. A metal lens 16 has the advantage of being flat, extremely thin, very light, and compact, which greatly facilitates its integration into an optical device 14.

[0027] The metal lens 16 is configured to diffract the light beam F produced by at least one light source 12 so as to project at least one pattern, such as a set of dots, onto the face 5 of the driver of a motor vehicle 100. Thanks to its small size, a metal lens 16 can be easily integrated into a optical device 14 while maintaining high-quality optical properties. In other words, the optical properties of the meta-lens 16 allow the diffracting of the infrared light beam F that passes through it, and the small dimensions of this meta-lens 16 allow the miniaturization of the optical device 14 in order to facilitate its integration into the facial recognition system 10.

[0028] Furthermore, the properties of the meta-lens 16 allow for a wider field of view of the optical device 14. Thus, despite its small size, the wide field of view of the optical device 14, which includes a meta-lens 16, is able to cover a greater area than known prior art optical devices. This is illustrated in particular in Figures 2 and 3: in [Fig. 2], the pattern projected onto the driver's face using a known prior art device extends over a narrower field of view, while in [Fig. 3], the pattern projected onto the driver's face using the optical device 14 in the facial recognition system according to the invention extends over a wider field of view, which makes it possible to obtain better contrast between the driver's face and their environment, such as the passenger compartment. A "wide" field of vision is defined as a field of vision greater than or equal to 120° x 20°.

[0029] More specifically, the meta-lens 16 of the optical device 14 comprises a nanostructure configured to diffract the light beam F that passes through this meta-lens 16. Here, "nanostructure" means a surface with nanometer-sized features. These features are generally produced by nanostructuring, that is, by electronic, ionic, or optical etching of thin films arranged in periodic lattices. An example of a nanostructure of a meta-lens 16 is illustrated in particular in [Fig. 4].

[0030] The nanostructure of the meta-lens 16 makes it possible to obtain a wide field of view without being affected by stray light from unwanted diffraction orders. For example, the nanostructure of the meta-lens 16 can be configured to obtain, in particular, five, seven, or even nine diffraction orders. This is achieved by parameterizing the geometry of the nanostructure of the meta-lens 16 according to the wavelength of the light from the light source 12.

[0031] The nanostructure of the meta-lens 16 may, in particular, include an arrangement of nanopillars 17 on its surface. More specifically, the nanostructure may comprise a quasi-periodic array of nanopillars 17. By "quasi-periodic," it is meant here that the nanopillars 17 within the nanostructure are placed at more or less regular intervals from each other. The aim is then to optimize the geometry of the nanopillars 17 of the nanostructure of the meta-lens 16 in order to project the light from the light source 12 over a wide field of view.

[0032] By way of example, the nanopillars 17 of a nanostructure can be placed at intervals of 650nm from each other. In other words, the centers of two neighboring nanopillars are 650nm apart.

[0033] According to one example, the nanopillars 17 comprise a material with a refractive index greater than 2, and more particularly greater than 3 or more particularly equal to 3.47. All the nanopillars 17 of the metal-lens 16 have the same refractive index. The nanopillars 17 are surrounded by air.

[0034] The nanopillars 17 are, for example, placed on a transparent substrate. The material of the transparent substrate is chosen to provide suitable structural support and to allow the majority of the light passing through it to pass through. Such substrate materials include, for example, fused silica, borosilicate glass, or rare-earth oxide-based glasses. As an example, the optical index of the transparent material chosen for the substrate is, for example, greater than or equal to 1.5.

[0035] In general, the nanopillars 17 can have cross-sections of different sizes. Indeed, the shape and size of the cross-section of a nanopillar 17 of the nanostructure have an impact on the propagation speed of the light beam F that passes through said nanopillar 17, thus modifying the optical properties of the wavefront of said light beam F that passes through the optical device 14.

[0036] The nanopillars 17 may in particular have a cylindrical shape; in this case, they may include a circular base, as illustrated in [Fig.4] and [Fig.5].

[0037] Furthermore, these nanopillars 17 can have the same height configured to modify a phase of the light beam F produced by the light source 12. For monochromatic light having a given wavelength, the ideal height of the nanopillars 17 can be determined to obtain the highest possible transmission rate.

[0038] According to one embodiment, the nanopillars 17 have a size less than or equal to the wavelength of the light from the light beam F. In this way, the nanopillars 17 can diffract the light beam F produced by at least one light source 12 and passing through the metal lens 16.

[0039] The distribution of the nanopillars 17 within the nanostructure of the meta-lens 16 can be optimized to contribute to a change in the direction of propagation of the light beam. In other words, the distribution of the nanopillars 17 within the nanostructure can be specifically configured to influence the direction of propagation of the light beam F in order to modify the orientation of the light beam at the exit of the meta-lens 16.

[0040] More generally, a meta-lens 16 comprises a quasi-periodic array of nanopillars 17 having different diameters so as to influence the speed of light propagation through the nanopillars 17 so as to modify the properties of the wavefront of the light beam F which passes through the metal lens 16. Indeed, for two nanopillars 17 having the same height and different diameters, the nanopillar with a larger diameter will slow down the propagation of the light which passes through it more than the nanopillar with a smaller diameter.

[0041] The diameter of the nanopillars 17 individually influences the propagation speed of the beam F passing through them, which can be used for diffraction of the light beam F in order to project a set of points into the passenger compartment 1 of the motor vehicle and more particularly onto the face 5 of a driver of said vehicle. According to one embodiment, the diameter of the nanopillars 17 is between 100 nm and 500 nm.

[0042] Thus, the distribution of the nanopillars 17 within the nanostructure of the meta-lens 16 makes it possible to diffract the light from the light source 12 uniformly towards the targeted diffraction orders with high efficiency. This is illustrated more particularly in [Fig. 5].

[0043] The point set may comprise at least two distinct light points, preferably at least five distinct points, or even several dozen distinct points. The light points may be aligned in a single line or arranged in a matrix to form several lines and / or columns of light points. Each point may be intended to illuminate a particular area of ​​the face, said area being chosen from among the eyes, nose, mouth, chin, cheekbones, or ears, or even any other characteristic point of the face 5 of a vehicle user 100. Some points of the point set may also be projected beyond the driver's face 5, i.e., projected into an environment close to the driver's face 5. In this way, an effective distinction can be made between the driver's face 5 and its surroundings.

[0044] To accurately extract a depth map of the scene illuminated by the light source 12, it is preferable that each projected light point have sufficient brightness. Thanks to the nanopillars 17 of the nanostructure of the metal-lens 16, the intensity of all the light points is similar, or even identical, even for a higher diffraction order. Uniform and / or homogeneous intensity for all the light points thus projected over a wide field of view makes it possible to obtain more precise images, which can allow for a more detailed and therefore more relevant analysis.

[0045] Alternatively, the projected pattern could be something other than a set of points. The pattern could comprise a set of segments and / or curved lines forming a geometric pattern. By way of example, this geometric pattern could include an elliptical shape such as a circle or a polygonal shape. The pattern could also comprise a network of segments forming a grid.

[0046] In the facial recognition system 10, the light-receiving device 18 includes, for example, a sensor configured to take pictures 25 of the driver's face 5 when it is illuminated by the light source 12. In addition, the light-receiving device 18 includes at least one analysis unit 19 configured to perform facial recognition from the light reflected by the face 5 of the motor vehicle driver. Thus, the analysis unit 19 analyzes at least one picture 25 taken by a sensor of the light-receiving device 18 to extract biometric markers enabling the identification of the person who was the subject of the picture 25. This corresponds to step 11 illustrated in [Fig. 6].

[0047] The light-receiving device 18 is fixed to a structure or a non-moving element of the passenger compartment 1, such as the windshield, the central rearview mirror, or the body pillar. In this way, the shots 25 fully include the face 5 of a driver, regardless of their size or seat adjustment, when the driver is seated in the driver's seat.

[0048] The analysis unit 19 is configured to perform facial recognition from the light reflected by the face 5 of the driver of the motor vehicle by extracting biometric markers. Thus, the analysis unit 19 is, for example, configured to analyze the reflections present in the photographs 25 in order to detect blood pulse, skin texture, or even the geometry of a face 5. Alternatively, the analysis unit 19 can be configured to interpret the received light as eye or facial movements 5 of the driver.

[0049] The memory unit 20 is configured to store at least one personalized driver profile. Here, "personalized profile" refers to a set of biometric markers that constitute a unique identifier for a person based on their facial geometry 5 and facial expressions. The memory unit 20 includes, for example, a database that lists the biometric markers of one or more people.

[0050] The processing unit 22 is configured to compare the biometric markers extracted from a photograph 25 taken by the sensor of the light receiving device 18 with the biometric markers of each person already registered in the memory unit 20. This comparison makes it possible, for example, to determine a match score between the biometric markers extracted from a photograph 25 and the biometric markers listed in a personalized driver profile. This corresponds to step E2 illustrated in [Fig. 6].

[0051] When the matching score between the biometric markers extracted from a shot taken 25 by the sensor of the light receiving device 18 and the biometric markers of a personalized profile stored in the memory unit 20 is above a minimum threshold, the facial recognition system 10 determines the identity of the person depicted in the image 25 taken by the sensor of the light receiving device 18 as being that of the given person. This corresponds to step E3 illustrated in [Fig.6].

[0052] The executive unit 24 is configured to perform an action on at least one parameter of the vehicle 100 based on a result from the comparison performed by the processing unit 22. According to a particular embodiment and in the case where the matching score is greater than a minimum threshold, the executive unit 24 can perform one or more actions such as a personalized welcome for the driver allowing, for example, to configure settings of the vehicle 100 according to the driver's preferences, in particular the adjustment of the position of the driver's seat and / or the inclination of the mirrors in the rearview mirrors, or allowing, for example, in addition or alternatively, the starting of the vehicle, in order to automatically start the vehicle by facial recognition of the driver.

[0053] According to another embodiment and in the event that the matching score is less than a minimum threshold, the executive unit 24 can perform one or more actions such as locking the steering wheel and / or the gear lever and / or the pedals to prevent the person sitting in the driver's seat from taking possession of the vehicle 100. Alternatively or in addition, the executive unit 24 can trigger an alarm and / or alert the owner of the vehicle by means of communication, such as an alert message sent on a portable electronic medium for example.

[0054] Preferably, the facial recognition system 10 can provide differentiated processing depending on the application, in order to take into account the level of security required by each action. For example, the vehicle start application requires highly secure facial recognition processing to prevent vehicle theft, whereas fraudulent use of the facial recognition system 10 associated with the personalized driver welcome application will have limited consequences.

[0055] Thus, a method for operating a facial recognition system 10 to identify a person from one or more photographs 25 of their face 5 may include the following steps: emission of a light beam F in the infrared range from at least one light source 12 through at least one optical device 14 comprising a metal lens 16 and in the direction of a face 5 of a driver of a motor vehicle; diffraction of said light beam F by the metal lens 16 of the optical device 14 so as to project at least one pattern onto the face 5 of the driver; reception of the light reflected by the face 5 of the driver in the light receiving device 18; execution of facial recognition from said light by the unit analysis 19; comparison of the facial recognition result with at least one personalized driver profile previously recorded in memory unit 20 by processing unit 22, implementation of an action on at least one parameter of vehicle 100 based on a result from the comparison by at least one executive unit 24.

[0056] Optionally, the facial recognition system 10 of a motor vehicle driver's face 5 can be used to monitor driver alertness, particularly on long and / or monotonous journeys. By continuously analyzing the driver's facial expressions, eye movements, number of blinks, head position, etc., the facial recognition system 10 can detect signs of fatigue, distraction, and / or drowsiness. In such cases, the executive unit 24 can issue a warning in the form of an audible and / or visual signal to draw the driver's attention and alert them to their condition.

[0057] Such a process can be carried out using a computer program product which includes program code instructions recorded on a computer-readable medium to implement the steps of said process when said program is running on a computer.

[0058] Furthermore, a data recording medium, readable by a computer and on which is recorded the computer program including program code instructions for implementing the method described above, can also be considered.

[0059] Thanks to a lighter and less bulky facial recognition system, personalized profiles of motor vehicle drivers and / or proactive safety procedures can be put in place, which can contribute to a safer and more secure driving experience.

Claims

Claims

1. System (10) for facial recognition of a face (5) of a driver of a motor vehicle (100), the facial recognition system (10) comprising a light source (12) configured to produce a light beam (F) in the infrared range and an optical device (14) comprising a meta-lens (16), the optical device (14) being arranged on the path of the light beam (F) produced by the light source (12), the meta-lens (16) being configured to diffract the light beam (F) so as to project at least one pattern in a field of vision and in particular onto a face (5) of a driver of a motor vehicle (100).

2. Facial recognition system (10) according to claim 1, characterized in that it further comprises: - at least one light receiving device (18) comprising at least one analysis unit (19) configured to perform facial recognition from the light returned by the face (5) of the driver; - at least one memory unit (20) configured to store at least one personalized profile of a driver; - at least one processing unit (22) configured to compare the results of the facial recognition performed by the light receiving device (18) with the at least one personalized driver profile recorded in the memory unit (20), - at least one executive unit (24) configured to perform an action on at least one parameter of the vehicle (100) as a function of a result resulting from the comparison performed by the processing unit (22).

3. Facial recognition system (10) according to one of the preceding claims, characterized in that the pattern intended to be projected onto a face (5) of a motor vehicle driver by the optical device (14) comprises a set of light points having a uniform intensity.

4. Facial recognition system (10) according to one of the preceding claims, characterized in that the meta-lens (16) comprises at least a nanostructure comprising a transparent substrate with an arrangement of nanopillars (17) on its surface and in that the nanopillars (17) have a size less than or equal to the wavelength of the light of the light beam (F) for diffracting the light beam (F) produced by the at least one light source (12) and passing through the meta-lens (16).

5. Facial recognition system (10) according to the preceding claim, characterized in that the nanopillars (17) have a cylindrical shape and characterized in that the diameter of the nanopillars (17) is between 100nm and 500nm.

6. Facial recognition system (10) according to claim 4 or 5, characterized in that the nanopillars (17) comprise a material with an optical index greater than 2, and more particularly greater than 3 or more particularly equal to 3.

47.

7. Method of operating a system (10) for facial recognition of a face (5) of a motor vehicle driver, the method comprising the following steps: - emitting a light beam (F) in the infrared range from the at least one light source (12) through the at least one optical device (14) comprising a meta-lens (16) and in the direction of a face (5) of a motor vehicle driver; - diffraction of said light beam (F) by the meta-lens (16) of the optical device (14) so as to project at least one pattern onto the face (5) of the driver; - receiving the light returned by the face (5) of the driver in the light receiving device (18); - performing facial recognition from said light by the analysis unit (19);- comparison of the result of the facial recognition with at least one personalized driver profile previously recorded in the memory unit (20) by the processing unit (22), - carrying out an action on at least one parameter of the vehicle based on a result from the comparison by at least one executive unit (24).;

8. A computer program product comprising program code instructions recorded on a computer-readable medium for implementing the steps of the method according to the preceding claim when said program is running on a computer.

9. A computer-readable data storage medium on which is recorded a computer program comprising program code instructions for implementing the method according to claim 7.

Images