DEVICE AND METHOD FOR ACQUIRING ON-THE-GO DERMATOGLYPHS USING MULTIPLE CAMERAS

Abstract

DEVICE AND METHOD FOR ON-THE-GO DERMATOGLYPH ACQUISITION USING MULTIPLE CAMERAS The present invention relates to a non-contact device for acquiring dermatoglyph images of a part of a hand in translational motion, by acquiring images of said hand part (D) at several different times (ta, tb, tc) of the translational motion, some images being acquired by a first lateral camera (CL1) along a first optical axis (A21) and others by a second lateral camera (CL2) along a second optical axis (A22), said optical axes (A21, A22) intersecting (I) upstream of the acquisition plane (P), said images acquired at several times (ta, tb, tc) corresponding to different positions of said hand part (D) within the acquisition volume (10), and at least two of said images, acquired along the different optical axes, image a common area of the hand part (D). Figure for the abbreviation: Fig. 3

Description

Title of the invention: DEVICE AND METHOD FOR ACQUIRING ON-THE-GO DERMATOGLYPHS USING MULTIPLE CAMERAS

[0001] The invention relates to the field of on-the-fly dermatoglyph image acquisition devices, for the acquisition of a biometric imprint of a part of a user's hand such as the palm of the hand, fingers of the hand, or the whole hand.

[0002] It is known from the prior art of contactless biometric fingerprint acquisition devices capable of acquiring a dermatoglyphic image of an individual's finger or palm. The problem with contactless devices is the need to acquire several images simultaneously from different perspectives to reconstruct an edge-to-edge representation of a dermatoglyphic image of the hand part (for example, a scroll), which implies a multiplication of optical imaging means, their synchronization, and a lack of compactness in such devices.

[0003] The invention aims to solve the aforementioned problems of the prior art by proposing an efficient and simple biometric fingerprint capture device, suitable for multi-finger acquisition and enabling the reconstruction of an edge-to-edge representation of at least one dermatoglyph of a part of the hand while limiting the complexity of the device, the part of the hand being able for example to include several fingers and as such several dermatoglyphs or to include a palm and as such a single but extended dermatoglyph.

[0004] The invention relates to a non-contact device for acquiring dermatoglyphic images of at least a part of a user's hand, said acquisition device comprising: - an acquisition volume extending on either side of an acquisition plane and constituting a free space area forming a passage for said part of the hand in translational motion along a lateral direction parallel to the acquisition plane; and - an electronic compartment, said electronic compartment comprising a fixed imaging system, including at least two lateral cameras, each having its own acquisition field partially covering the acquisition volume, said acquisition fields being distinct with an overlap, defined as an intersection surface on a union, in the acquisition plane of 0 to a maximum of 50%, said lateral cameras having intersecting optical axes intersecting upstream of the acquisition plane, characterized in that: each of said side cameras is configured to acquire an image of said part of the hand, each of the side cameras acquiring said image at a different instant of the lateral translation movement of said part of the hand in the acquisition volume, each instant being specific to each side camera, said images acquired at said instants forming a series of images comprising a plurality of views of said part of the hand corresponding to different positions of said part of the hand in the free space of the acquisition volume and imaging a common area of ​​said part of the hand.

[0005] The device according to the invention, by virtue of its lateral cameras having optical axes intersecting upstream of the acquisition plane, i.e., converging with intersection before the capture zone, makes it possible to obtain a series of images corresponding to a scroll at different positions of said hand part during the translational movement, so that the images complement each other and can allow edge-to-edge representation, their fusion being possible thanks to the existence of the common imaged area included in said images, at least two at a time, the common imaged area being in particular a dermatoglyphic part of the hand part. The device according to the invention makes it possible in particular for a "left" lateral camera to image the right part of the acquisition volume and a "right" lateral camera to image the left part of the acquisition volume,The optical axes of the lateral cameras are advantageously inclined in opposite directions to the direction of hand movement, and by design, the entire hand portion does not need to be included in all acquisitions, ensuring compactness.

[0006] Preferably, the overlap of said distinct acquisition fields in the acquisition plane is between 30% and 40%, in particular 34%, which allows for a compact construction, particularly in an embodiment without a central camera.

[0007] In one embodiment of the invention, at least one part of the hand comprises at least two fingers, preferably four fingers, or a palm, the common area being specific to each finger or palm respectively, which allows a diversity of applications to fingerprints, including multi-finger prints, as well as palm prints, taking into account the writer's palm.

[0008] Advantageously, each finger of the hand part is seen in its entirety in the image acquired at the acquisition instant specific to each lateral camera, which ensures an acquisition of the phalange.

[0009] In one embodiment of the invention, the imaging system comprises a central camera whose acquisition field covers all or part of the fields acquisition of the side cameras in the acquisition plane, and preferably covering the entire acquisition volume,

[0010] This optional central camera provides a contextual view and allows the finger's movement speed to be determined by tracking the hand's movement frame by frame, thanks to its wider field of view. The central camera is positioned, in particular, such that an intersection at the acquisition plane of a projection of an optical axis of the central camera onto a plane defined by the optical axes of two of the lateral cameras is located between the intersections of the optical axes of said two lateral cameras with the acquisition plane, or, in other words, the central camera is positioned such that, in a plane defined by the optical axes of two of the lateral cameras, the angle between: - a projection of an optical axis of said central camera - and a normal to the acquisition plane is less than the angles between said normal and each of the optical axes of said two lateral cameras in said plane as defined. It is on the basis of this geometric construction in said plane that this optional camera is called central, as opposed to lateral, although it is not necessarily centered in the strict sense.

[0011] Preferably, the overlap of said distinct acquisition fields in the Acquisition plan is worth 0% with central camera.

[0012] Advantageously, the optical axis of the central camera is not in the common plane of the optical axes of each of the lateral cameras, which makes it possible to improve the image of the fingertip.

[0013] In one embodiment of the invention, the non-contact acquisition device includes a processing unit connected to the imaging system, the processing unit being configured to fuse the acquired images imaging the common area, which makes it possible to determine an edge-to-edge representation of the dermatoglyph of the hand part.

[0014] Advantageously, each camera comprises a single sensor and a single lens.

[0015] In one embodiment of the invention, the angular opening of the fields The acquisition angle of each of the side cameras is less than 25°, specifically less than 15° and preferably approximately 10°. This offers advantages in terms of compactness compared to wide-field side cameras, as well as cost, since the lens is simpler for the same performance with a smaller object area. It also improves image quality because the resolution is increased for the same number of pixels on the sensor, and finally, it reduces size because a narrower field of view allows for a shorter focal length and therefore less extension. In particular, in the absence of a central camera, the maximum angular aperture of the The acquisition fields of the side cameras are preferably 25° and in the presence of a central camera it can be limited to 15°, with for example a minimum angular opening of 5°.

[0016] Advantageously, the side cameras have an acquisition frequency higher than an acquisition frequency of said central camera, which allows the device to be optimized because for the same quality, the quantity of data is greater with the wider field central camera.

[0017] In one embodiment of the invention, in a plane common to the optical axes of each of the lateral cameras, the optical axis of at least one of the lateral cameras is inclined with respect to a projection of a normal to the acquisition plane in said common plane between 10° and 60°, preferably between 25 and 45°, which makes it possible to dimension the common area, for example this makes it possible for a finger to obtain an imaged common area of ​​approximately one-third of the dermatoglyph.

[0018] In one embodiment of the invention, two of said side cameras are positioned symmetrically with respect to a cutting plane of the device, which allows for right-hand and left-hand acquisition symmetry, as well as simple implementation.

[0019] Advantageously, the plane common to the optical axes of each of the lateral cameras is inclined with respect to a normal to the acquisition plane, which allows for better visibility of the fingertips.

[0020] Advantageously, the acquisition device includes a movement guide for said part of the user's hand, in particular by means of mechanical stops, constituted for example by the notched shape of the lateral sides of the user compartment thus designed to encourage the lateral translation movement of the user's hand, or by means of a guidance interface, which allows by means of obstacles providing a free space sufficiently small to constrain the part of the hand to be in the acquisition volume, or by the display and / or emission of information to the user to help him to carry out the translation movement of his hand and encourage him in particular to spread his fingers so as to improve the acquisitions of the edges of the fingers.

[0021] In one embodiment of the invention, the electronic compartment includes a lighting means suitable for illuminating the acquisition volume, said lighting means comprising at least two light sources, each of which has a mean diffusion axis substantially collinear with the optical axis of each camera, which allows the acquired hand part to be uniformly illuminated.

[0022] There may be one less light source than the number of cameras, to optimize the footprint by placing a source between two cameras.

[0023] Preferably the lighting means comprises as many light sources as there are cameras in the acquisition device. In one embodiment of the invention, the electronic compartment includes a lighting means suitable for illuminating the acquisition volume, said lighting means comprising a plurality of light sources, arranged annularly around each camera or between the cameras, the first alternative allowing good control of the lighting, the sources being preferably mostly close (for example offset laterally a few centimeters above) to each camera so as not to have an unlit area with optionally sources arranged on the sides of the electronic compartment to better illuminate the sides of the finger, and avoid their dark appearance resulting from their inclination, and the second alternative allows limiting the bulk by placing all the sources between the cameras, with preferably a subset placed less than 10 cm from the cameras,particularly in an area describing a dimension that is larger longitudinally (front / back) than laterally (left / right) so as to illuminate along the finger. Advantageously, the light beam is active for a duration of less than 2 ms per image acquisition, preferably less than 0.5 ms; this flash-like embodiment avoids the motion blur inherent in moving objects.

[0024] Advantageously, each camera of the imaging system has a depth of field extending on both sides of the acquisition plane, the sensor of each camera having an acquisition speed of at least 8 frames per second, and preferably at least 15 frames per second, which makes it possible to acquire at least three images of said part of the hand in translational motion at a predetermined maximum speed of movement during its passage through the acquisition volume.

[0025] Advantageously, the lighting means emits red or near-infrared light, i.e., with a wavelength greater than 600 nm, which allows, in particular, by diffusion within the hand area, for the entire hand area to be well illuminated without the need for fine-tuning the position of the lighting means. Alternatively, the lighting means emits light at a wavelength less than 580 nm, in particular blue or green or near-ultraviolet light, which preserves the visibility of the details of the papillary pattern, as this lighting penetrates only superficially into the hand area.

[0026] Advantageously, the acquisition plane is inclined relative to a horizontal plane, particularly for ergonomic reasons.

[0027] Advantageously, the acquisition device includes at least one polarizer disposed in front of the imaging system or the lighting means. In particular, in the case of polarizers in front of the lighting means and in front of the imaging system, these polarizers are linear and oriented in the same direction – or if they are circular with a symmetrical direction of rotation between the one in front of the imaging system and the one in front of the illumination. Preferably, a single linear polarizer is arranged in front of the imaging system and the illumination means.

[0028] Advantageously, the acquisition volume extends laterally sufficiently to image a finger, 4 fingers of a hand simultaneously and / or the palm, covering respectively at least 3 cm of the acquisition plane to image a finger, at least 7 cm, preferably at least 7.8 cm of the acquisition plane to image the 4 fingers simultaneously or covering at least 10 cm of the acquisition plane, preferably between 10 and 16 cm, to image the palm.

[0029] Advantageously, the device includes a three-dimensional information acquisition system for the acquisition volume such as a three-dimensional time-of-flight camera, or a stereovision camera system, in particular taking advantage of the cameras of the imaging system, or a three-dimensional structured light camera.

[0030] According to another aspect of the invention, a method is proposed for non-contact acquisition of dermatoglyphic images of at least a part of a user's hand, said hand part traversing an acquisition volume in a translational movement in a lateral direction parallel to an acquisition plane, said method comprising the acquisition of images of said hand part at a plurality of different instants of the translational movement of said hand part, said images acquired at different instants forming a series of images comprising a plurality of views of said hand part corresponding to different positions of said hand part in the acquisition volume at a plurality of instants of the movement and at least two of said images are acquired along a different optical axis,said optical axes intersecting upstream of the acquisition plane and said at least two of said images imaging a common area of ​​said hand part. This method presents the same advantages as the device according to the invention, allowing for different images of the hand part but imaging at least two by two a common area of ​​the hand part and in which there is therefore an overlapping area of ​​the hand part so that the images can complement each other for fusion, said images of the hand part being acquired along at least two different optical axes.

[0031] In one embodiment, the process is implemented by an acquisition device according to the invention, so that the at least two of said images, acquired along different optical axes, at a plurality of times, comprise at least one image acquired by one of the side cameras and one image acquired by the other side camera at two different times.

[0032] In one embodiment, said at least one part of hand comprises at least two fingers, preferably four fingers, or a palm, the common area being specific respectively to each finger or palm.

[0033] In one embodiment, the method according to the invention includes a step of segmenting the acquired images by finger.

[0034] In one embodiment, the method according to the invention includes a step of reconstructing a fused image of a surface to be represented of said part of the hand from the acquired images.

[0035] In one embodiment, the step of reconstructing the fused image from the acquired images comprises: - an extraction of characteristic points specific to said hand part in each image of the plurality of views; - a matching of the common characteristic points between each pair of views of the plurality of views; - an assembly of the plurality of dermatoglyph images to join parts of said images according to the correspondences determined in the previous step; - a determination, from the plurality of assembled dermatoglyph images, of an edge-to-edge representation of the dermatoglyph of the hand part.

[0036] It may be provided that the edge-to-edge representation of the dermatoglyph of the hand part can be displayed on the GUI 605 user interface of the device 1 if it has a screen or on a connected remote screen or input / output module 607 or via the network interface.

[0037] Advantageously, the method includes a step of selecting images from among the acquired images, which makes it possible to apply the steps of the method to said image selection and not to all of them, particularly in the case of image acquisition by cameras whose output is a video stream.

[0038] Advantageously, a wide-field image is acquired by the central camera at substantially the same time as the acquisition time specific to each of the selected images acquired by one of the side cameras, which allows the wide-field image to be used to assist image fusion.

[0039] Advantageously, the method includes a user guidance step prompting the user to perform a translational movement of said hand part in the acquisition volume.

[0040] Advantageously, an additional image of the hand part is acquired by one of the cameras, lateral or central, during an insertion movement of the hand perpendicular to the translational movement, in particular prior to the translational movement, which makes it possible to obtain, for example, an image of 4 fingers and / or of the palm, and this additional image can also be used in the reconstruction of the merged image, or simply to confirm that the palm and finger images acquired during the scan movement(s) belong to the same hand, or to confirm that the images acquired during the scan movement(s) all correspond to fingers of the hand acquired in the additional image.

[0041] Advantageously, a computer program is proposed comprising instructions adapted to the implementation of each of the steps of the process according to the invention when said program is executed on a computer.

[0042] Advantageously, a non-transient means of storing information, removable or not, partially or totally readable by a computer or a microprocessor, comprising code instructions of a computer program for the execution of each of the steps of the process according to the invention is proposed.

[0043] The attached drawings are given by way of non-limiting examples:

[0044] [Fig.1] represents, according to a schematic three-dimensional view, a contactless acquisition device according to an embodiment of the invention;

[0045] [Fig.2] represents, according to a schematic three-dimensional view, the previous non-contact acquisition device in use;

[0046] Figures 3 and 4 represent, according to a schematic diagram, a cross-sectional view along x,z of the architecture of the contactless acquisition device according to an embodiment of the invention, without a central camera;

[0047] [Fig.5] represents, according to a schematic diagram, a schematic longitudinal cross-sectional view of the architecture of the non-contact acquisition device according to an embodiment of the invention with a central camera;

[0048] [Fig.6] represents a schematic block diagram of an information processing unit for the implementation of one or more embodiments of the invention;

[0049] Figure 7 illustrates the acquisition process in one embodiment of the invention; and

[0050] Fig. 8 illustrates a plurality of views acquired by the device according to the invention.

[0051] Identical references will be used from one figure to another to designate identical or similar elements, in their form or in their function.

[0052] In the figures, the x, y, z axes are the axes taken in the reference frame of the non-contact acquisition device, that is to say that the Ox axis, or lateral axis, is the left-right axis of the device, collinear with the translation axis of the hand part, that the z axis is perpendicular to the acquisition plane and that the y axis, or longitudinal axis, designating the front-back axis of the device, is perpendicular to the x and z axes.

[0053] For the sake of brevity, the term "approximately" refers to values ​​within ±5% or 5 degrees when referring to angles. The edge-to-edge representation of a dermatoglyph implies that, for a finger, the surface to be represented is all or part of the surface extending from a first nail edge to a second nail edge via the pulp of each finger and respectively for a palm the surface to be represented is the surface extending from a first edge of the palm to a second edge of the palm via a palmar face of the hand.

[0054] The term side camera refers to a camera arranged in such a way as to acquire dermatoglyphs extending over the edges of the hand part.

[0055] The common area of ​​the hand part refers in particular to a common area of ​​dermatoglyph of the hand part.

[0056] Fig. 1 illustrates a non-contact acquisition device 1 capable of acquiring dermatoglyph images of at least part of a user's hand, for example the palm of a hand, one or more fingers of a hand, or the entire hand.

[0057] The contactless acquisition device 1, whose reference frame x, y, z is shown, comprises: - a frame comprising an electronic compartment 20 and a user compartment, - the electronic compartment 20 comprising a fixed imaging system, - the user compartment of the frame defining here an acquisition volume 10 extending on either side of the acquisition plane P and being arranged to provide a free space area which contains the acquisition volume 10, said free space forming a passage for the part of the hand in translational motion along a lateral direction parallel to the acquisition plane.

[0058] The frame may comprise a single substantially parallelepiped-shaped block, also called a terminal, having walls common to the user compartment and the electronic compartment 20 or may be an assembly of a physical block of user compartment and a different physical block of electronic compartment.

[0059] The user compartment and the electronic compartment 20 can be separated from each other, for example by a partition glass. The partition glass is preferably located in the electronic compartment 20, particularly on its upper surface. The partition glass can advantageously be oriented or offset from the boundary between the two compartments (lowered, as illustrated in [Fig. 5] by reference numeral 16) to limit reflections or facilitate cleaning. The partition glass advantageously has optical functions, such as polarization or wavelength or angle filtering.

[0060] The user compartment includes at least one opening for the passage of the part of the hand whose dermatoglyph, i.e. the set of digital or palmar grooves which constitute the biometric impressions: digital or palmar, is to be imaged.

[0061] For example, as shown here, the side sides of the user compartment are notched, i.e., their edge is hollowed out, forming two lateral openings 11, to allow the passage in lateral translation along the x-axis of the user's hand.

[0062] According to an embodiment illustrated in [Fig. 1], the upper side of the user compartment corresponds to a visor 15, and the acquisition volume 10 is notably delimited by the walls of the user compartment. However, the compartment may be partially or entirely external to the frame, and the acquisition device may also be more open and not include a clearly delimited user compartment; for example, it may not include a visor. In this case, the upper limit of the acquisition volume is geometrically delimited by the depth of field of the imaging system, as may be the lower limit of the acquisition volume.In this implementation mode, the acquisition plane P is advantageously materialized by patterns (for example a line) on a wall (for example the wall at the back of the user compartment when it exists or an external wall to the device on which, for example, the frame is fixed) and / or a line projection, and / or an image on a screen of the guidance device.

[0063] The user compartment, which here delimits the acquisition volume 10, includes a front opening 11 and has a depth allowing the insertion of the user's hand from one fingertip, and in particular up to the palm, along the y-axis in a so-called longitudinal direction. In [Fig. 1] it is substantially parallelepiped in shape.

[0064] Without limitation, the height (z-axis) of the acquisition volume is constant in both the transverse (x-axis) and longitudinal (y-axis) directions, specifically from 4 cm to 10 cm to allow passage of the hand, including the entire hand, without hindering translational movement. The acquisition volume extends laterally (x-axis) sufficiently to image a finger, four fingers of a hand simultaneously, and / or the palm, covering at least 3 cm of the acquisition plane to image a finger, at least 7 cm, preferably at least 7.8 cm, of the acquisition plane to image the four fingers simultaneously, or covering at least 10 cm of the acquisition plane, preferably between 12 and 16 cm, to image the palm. The depth (y-axis) of the acquisition volume is, for example, 15 cm, so as to be able to image several fingers or a palm, and in particular the lateral part of the palm on the side opposite the thumb, also called the writer's palm.

[0065] In [Fig. 1], the acquisition volume has a median acquisition plane P that is substantially horizontal. This example is not limiting; depending on ergonomic choices, this acquisition plane P can be inclined with respect to a horizontal plane, for example, following the installation height of the non-contact acquisition device 1, if the acquisition volume 10 is positioned lower than the bend, its median acquisition plane P is advantageously inclined downwards in the depth direction relative to the longitudinal direction; and conversely if the acquisition volume 10 is positioned higher than the bend, its median acquisition plane P is advantageously inclined upwards.

[0066] In an alternative embodiment, the acquisition plane P could be in a vertical plane and the translational movement also.

[0067] Advantageously, the user compartment 20 includes a front opening 11 allowing the passage of a hand from the tips of the fingers and in particular to the beginning of the user's wrist, in a so-called longitudinal direction.

[0068] The non-contact acquisition device 1 of the invention can be used for an acquisition of a part of a hand M comprising one or more fingers D, for example an acquisition of a finger D, of several fingers D, and / or of a palm. The acquisition volume 10 is adapted to receive said user hand part M such that the direction of said user hand part M in the middle of the crossing of the acquisition volume during its translational movement corresponds to the direction of the orientation vector V, collinear with the y-axis, for example the fingers of the hand part M pointing in the direction of said orientation vector V.

[0069] According to a mode of use illustrated in [Fig.2], the acquisition by fixed imaging system of the non-contact acquisition device 1 is carried out while the user passes his hand part through the acquisition volume 10 of the user area in a translational movement in a lateral direction parallel, or substantially parallel, to the acquisition plane.

[0070] In this example, the movement of the hand M to enter and exit the user area follows a motion vector Ml, M2 of direction collinear with the lateral axis x, in the acquisition plane P. However, the translational movement can be unidirectional Ml or M2.

[0071] Figure 3 illustrates the architecture of the contactless acquisition device 1 according to an embodiment of the invention, in the xz plane, the device comprising: - an acquisition volume 10 extending on either side of an acquisition plane P and constituting a free space area forming a passage for said part of the hand D in translational movement along a lateral direction parallel to the acquisition plane P (along x), - an electronic compartment 20, the electronic compartment comprising a fixed imaging system, including at least two lateral cameras CL1, CL2, each having its own acquisition field Chl, Ch2 partially covering the acquisition volume 10, said acquisition fields being distinct with a overlap, defined as intersection surface on union, in the acquisition plane P from 0 to a maximum of 50% (preferably 34%), said lateral cameras CL1, CL2 having secant optical axes A21, A22 intersecting upstream I of the acquisition plane P, i.e. in convergence with overlap before the capture zone, each camera CL1, CL2 of the imaging system having a depth of field extending on both sides of the acquisition plane; And - a processing unit 106 connected to the imaging system 40 to process the images coming from it.

[0072] The device 1 therefore contains two lateral cameras arranged here, one on the left CL2 and the other on the right CL1 of the electronic compartment 20, such that the left lateral camera CL2 images the right part of the acquisition volume 10 and the right lateral camera CL1 images the left part of the acquisition volume 10. In this embodiment of the invention, the two lateral cameras are advantageously positioned symmetrically with respect to a section plane yz normal to the acquisition plane P of the device 1, which allows for simple implementation. Another lateral camera could be positioned so as to image fingertips, for example.

[0073] The side cameras CL1, CL2 of the imaging system each comprise at least one lens and one sensor, and in the case illustrated in the figure, each side camera comprises a single lens and a single sensor. The optical center 01, 02 of each of the side cameras CL1, CL2 is a geometrically predefined point of the imaging system, here defined according to the "pinhole" model, that is, as the point through which all rays passing from the imaged object to the sensor pass. The sensor comprises the photosensitive surface of the camera. The optical axis of each side camera A21, A22 is fixed in the coordinate system of the electronic compartment 20 and passes through the optical center 01, 02 specific to the side camera and the center of the sensor specific to the side camera. The optical axes A21, A22 are inclined with respect to the direction of hand movement in opposite directions.

[0074] In the figure, these optical axes A21 and A22 are in a vertical plane for illustrative purposes; however, this is not a limitation, as these optical axes can also be in a horizontal plane or advantageously inclined, depending in particular on an ergonomic choice resulting from the relative positions of the acquisition device 1 and the user. It is even advantageous for the plane common to the optical axes of each of the lateral cameras to be inclined with respect to a normal to the acquisition plane, which allows for better visibility of the fingertips.

[0075] The working distance of the imaging system, namely the distance between each objective lens and the center of the acquisition plane P along each optical axis, is predetermined and results from the choice of the lateral viewing angle [31,

[32] (here 30°) and the width of the acquisition volume (here 14 cm). In the embodiment of [Fig. 3], this distance is 35 cm. Alternatively, this distance could be variable, without affecting the invention.

[0076] In the embodiment shown in this figure, the two side cameras are configured to have an acquisition field partially covering the acquisition volume 10, each acquisition field preferably extending laterally over at least 60% of the width of the acquisition volume, i.e., extending here over at least 8.5 cm in width (along the x-axis) of the acquisition volume, which facilitates tracking the hand portion for fusion purposes. The angular aperture Q1, Q2 of the acquisition fields of each of the side cameras is less than 25°, preferably approximately 10°. However, overlap of the fields of the side cameras CL1, CL2 in the acquisition plane P is not necessary, as tracking of the hand portion can be achieved by other means.

[0077] In [Fig.3], for reasons of clarity, the part of the hand represented here is a single finger D. Each of said lateral cameras CL1, CL2 is configured to acquire an image of said part of the hand D at an instant of the lateral translation movement of the finger D from right to left in the acquisition volume 10, said instant being specific to each lateral camera CL1, CL2, said instants ta, te specific to each of said lateral cameras CL1, CL2 being different from each other (but not excluding that there is a common instant, here tb, different from ta and te but common to the two lateral cameras) and forming a plurality of instants ta, tb, te of the lateral translation movement of the part of the hand D in the acquisition volume 10, each of the acquired images imaging at least a common area of ​​the part of the hand D.

[0078] Considering a lateral translational movement of finger D in the acquisition plane P, arbitrarily from right to left, the solid black representation D(ta) of finger D corresponds to the first instant ta, the solid dotted pattern representation D(tb) of finger D corresponds to the second instant tb (here later than ta) and the empty contour representation D(tc) of finger D corresponds to the third instant te (here later than tb).

[0079] The three images acquired at the three times ta, tb, te thus image a common area of ​​the hand part, specifically here the lower central part of the dermatoglyph of finger D, which allows the processing unit 106 to fuse the acquired images to reconstruct an edge-to-edge representation of the dermatoglyph of finger D thanks to the overlap between the acquired images. The images acquired at the different times ta, tb, te form a series of images comprising a plurality of Views of finger D corresponding to different positions of finger D in the free space of the acquisition volume 10 (forming a passage for said part of the hand in translational motion (scanning), along a direction parallel to the acquisition plane) and at least two of said images acquired by said at least two lateral cameras CL1, CL2, at different times of the movement, image a common area of ​​finger D. Finger D is therefore imaged at least once by each lateral camera CL1, CL2 at a different position. Thus, by the configuration of device 1, the images acquired by the lateral cameras during the passage of finger D differ and complement each other; thus, at time ta only the image acquired by the lateral camera CL2 includes finger D, at time tb both lateral cameras image finger D which is in each of their acquisition fields, and at time te only the lateral camera CL1 images finger D.Each image acquired by each lateral camera within its field of view illustrates moments in the translational movement of the hand part, sequencing them chronologically to form the representative series of the movement. In other words, the representative series of the translational movement of the hand part is formed by images acquired by each of the lateral cameras, that is, along two optical axes.In the case of a right-to-left movement, the first images of the series will be mostly acquired by the left camera CL2 and then the last of the series will be mostly acquired by the right camera CL1, knowing that in the event of overlap of the acquisition fields of the side cameras CL1,CL2 in the acquisition plane P the intermediate images corresponding to a position of the part of the hand D in the acquisition field common to the two cameras will be acquired by one or the other or both side cameras, simultaneously or not (the side cameras do not require synchronization between them).The sensor of each side camera CL1, CL2 has an acquisition rate of at least 8 frames per second, and preferably at least 15 frames per second, which allows at least three images to be acquired of said part of the hand in translational motion at a predetermined maximum speed of movement during its passage through the acquisition volume.

[0080] The processing unit 106 is configured to fuse the acquired images imaging the common area, and in particular for each common area of ​​the hand part D, which makes it possible to determine an edge-to-edge representation of the dermatoglyph because the common area is included in each image, which also covers a larger surface area than just the common area, thus providing additional information about a part of the finger D (or palm) that another camera did not see, in addition to the imaged common area (one common area per finger, or per palm). Preferably, the information processing unit 106 receives and processes the images received from the imaging system as well as the acquisitions from the acquisition system. of three-dimensional information when present (not the case illustrated here). The information processing unit 106 typically comprises at least one calculator, computer, tablet, or other device enabling the execution of a computer program responsible for controlling the various stages of the process according to the invention. The information processing device 106 includes, in particular: - a control module for the imaging system which allows, in particular, the control of image acquisitions by the side cameras CL1, CL2; - a module for reconstructing a merged image of a surface to be represented of said part of the hand from the acquired images; these modules can be hosted locally in processing unit 106 of electronic compartment 20 or on a remotely accessible server.

[0081] Advantageously, the acquisition device includes a guiding element (not shown here) for the movement of said part of the user's hand, in particular by means of mechanical stops or by means of a guiding interface, which makes it possible to assist the user in helping him to carry out the translational movement of his hand, the guiding interface being able to be part of the processing unit 106.

[0082] The electronic compartment 20 includes a lighting means 60 capable of generating at least one light beam towards the acquisition volume, thus illuminating the hand part D. The lighting means 60 is used to illuminate the surface(s) to be represented on the hand part and comprises, for example, a single light source or a plurality of light sources. A light source may be an extended or point source. A light source 60 may have a wide or concentrated beam angle. The electronic compartment 20 here comprises two light sources, i.e., as many as there are cameras CL1, CL2, and each has a mean beam axis approximately collinear with the optical axis of each camera A21, A22, thus enabling uniform illumination of the acquired hand part.The light sources are positioned close to each camera CL1, CL2, specifically less than 10 cm from each, allowing for precise control of the lighting to avoid any unlit areas. Furthermore, they are located around the periphery of cameras CL1, CL2, positioning them on the sides of the electronic compartment 20. This allows for better illumination of the sides of the finger, preventing its dark appearance due to its tilt. Advantageously, the light beam is active for less than 2 ms per image acquisition, preferably less than 0.5 ms. This flash-like implementation prevents motion blur of the moving finger D. Without limitation, the multiple light sources are configured so that their combined coverage covers the entire acquisition volume 10, even if each source illuminates only a portion of it, in order to avoid... no loss of image quality including when finger D is at the edge of the acquisition volume.

[0083] In the illustrated embodiment, the lighting means emits blue light, which preserves the visibility of the details of the papillary pattern, this lighting only superficially penetrating the hand area.

[0084] The light sources of the lighting means 60 are here light-emitting diodes (LEDs) (or super-luminescent diodes SLDs).

[0085] The upper wall 15, also called the visor, preferably glazed, delimiting here the acquisition volume 10 of the user compartment, can be reduced, or even eliminated, when light leaks are neither dazzling nor dangerous, in particular in the case of infrared lighting.

[0086] The acquisition device here comprises a single polarizer (not shown) positioned in front of the imaging system and the illumination means 60. Thus, by using polarized illumination and an imaging system with the same polarization direction, the image of the surface is favored, as the light entering the finger D is depolarized by successive scattering, which is beneficial to the contrast between the ridges and furrows of the dermatoglyph. Preferably, this large polarizer is integrated into the separating glass 16, and then, if the latter is glazed, an orthogonal polarizer is optionally placed on the glazed portion of the upper wall 15 to prevent light leakage and protect the user from glare.

[0087] The acquisition device 1 could not include a visor (in the case of infrared light illumination in particular) and the line representing the acquisition plane P could then advantageously be drawn on the single vertical wall of the user compartment located at the bottom of said user compartment, so as to guide the user.

[0088] The cross-section of [Fig. 4] particularly illustrates the characteristic whereby the optical axes A21, A22 are inclined with respect to the direction of hand movement in opposite directions. Here, in the plane common to the optical axes of each of the lateral cameras CL1, CL2, the optical axis A21, A22 of at least one of the lateral cameras is inclined with respect to the projections of the normals (represented by dashed lines) to the acquisition plane P, each passing through the optical center of a camera 01, 02 in said common plane, forming the angles [31,

[32] with a value between 10° and 60°, preferably between 25° and 45°, in particular 30°, which allows an overlap, defined as the surface of the intersection on the union, of the acquisition fields of said lateral cameras in the acquisition plane of, for example, one-third.

[0089] Figure 5 represents, according to a schematic diagram, a cross-sectional schematic view of the architecture of the contactless acquisition device 1 according to another embodiment of the invention, in the y,z plane. In these figures the height (in z) of the acquisition volume 10 is not constant along the longitudinal direction (y), advantageously greater on the side of the front opening than at the bottom, so as to guide the movements of the user since the palm and wrist are thicker than the fingertips X.

[0090] The imaging system comprises a central camera CC with optical axis A2C passing through the optical center OC, and here passing (non-limitingly) through the fingertip X, such that the optical axis A2C coincides with the vector optical center OC - fingertip X. Each of the lateral cameras CL1, CL2, and the central camera CC has a depth of field extending on either side of the acquisition plane P. The central camera CC has a larger angular aperture QC than the lateral cameras Q1 and Q2; its acquisition field covers all or part of the acquisition fields of the lateral cameras CL1 and CL2 in the acquisition plane P, and here covers the entire acquisition volume. The two lateral cameras are aligned here on the x-axis, with the lateral camera CL2 obscuring the lateral camera CL1 (schematic view).In the embodiment shown in this figure, the two lateral cameras CL1, CL2 are configured to have an acquisition field partially covering the acquisition volume 10, each acquisition field preferentially extending laterally over at least 30% of the acquisition volume width, i.e., extending here over at least 5 cm in width (along the x-axis) of the acquisition volume, with an angular aperture Q1, Q2 of the acquisition fields of each of the lateral cameras less than 15°, here valued at 10°. In this embodiment, the acquisition field of the central camera CC has an angular aperture QC of 22° for an acquisition volume width of 14 cm, a focal length lens of 25 mm, and a magnification of 0.07, allows imaging of an area of ​​approximately 14 cm by 14 cm on a 1cm x 1cm sensor, capable of imaging an entire palm of the hand. The central camera CC is advantageously positioned between the lateral cameras, that is to say, in such a way that the intersection at the acquisition plane P of a projection of an optical axis A2C of said central camera CC in a plane defined by the optical axes A21, A22 of the two lateral cameras CL1, CL2 is located between the intersections of the optical axes A21, A22 of said two lateral cameras with the acquisition plane P. The central camera CC is positioned in such a way that in a plane defined by the optical axes A21, A22 of the two lateral cameras CL1, CL2, the angle between: . - a projection of an optical axis of said central camera - and a normal to the acquisition plane is less than the angles [31,

[32] between said normal and each of the optical axes A21, A22 of said two lateral cameras CL1, CL2 in said plane, and can to be null. This wider-field central camera CC allows for better tracking of the hand's movement, particularly determining its speed and facilitating three-dimensional reconstruction. Advantageously, the plane normal to the acquisition plane P, passing through the central camera CC, constitutes a plane of symmetry for the lateral cameras CL1 and CL2, thus enabling acquisitions of the same quality for the right and left hands. As seen here, the central camera is not necessarily centered within the surface of the electronic compartment defined by a section plane x, y, nor on an axis connecting cameras CL1 and CL2. Similarly, and advantageously, the optical axis A2C of the central camera is not in the plane common to the optical axes A21 and A22 of each of the lateral cameras, which improves the image of the fingertip D.

[0091] In the embodiment illustrated in this [Fig.5], and without limitation, the lateral cameras CL1, CL2 (which are aligned according to this view) have vertical optical axes A21, A22 and the central camera CC has an optical axis A2C inclined relative to the previous ones, in particular oriented so as to better image the dermatoglyphs present at the tip of finger D. This configuration could be reversed.

[0092] The electronic compartment 20 also includes a lighting means suitable for illuminating the acquisition volume 10, said lighting means comprising a plurality of light sources, arranged annularly around each camera CL1, CL2, CC, allowing good control of the lighting, the sources being preferentially placed close to the cameras (less than 10 cm).

[0093] Device 1 may also include a three-dimensional information acquisition system for the acquisition volume, such as a time-of-flight three-dimensional camera, a stereovision camera system (advantageously using the cameras already present in the imaging system), or a structured-light three-dimensional camera with a structured-light projector. The three-dimensional information acquisition system may be independent and include, for example, a time-of-flight three-dimensional camera, a stereovision camera system, or a structured-light three-dimensional camera, or it may use (alone or in addition to another camera) all or part of the cameras of the imaging system, for example, if one or all of the cameras of the stereovision system belong to the imaging system or if a sensor of the imaging system is also capable of acquiring structured light.Indeed, given that the imaging system includes by configuration multiple CL1, CL2 or even CC cameras, the three-dimensional information acquisition system is advantageously formed by a stereovision camera system, whose cameras already belong to the imaging system, especially since, in addition to the advantage of compactness, having a system calibrated between two cameras imaging a common area helps to. The matching of the common area between the two images, particularly for the purpose of merging. The three-dimensional information acquired by the three-dimensional information acquisition system can advantageously be used by the processing unit 106 to select the images, i.e., the respective acquisition times, as well as for the step of reconstructing a merged image of a surface to be represented of said hand part from the acquired images. Preferably, it is optimal for the three-dimensional information acquisition system to cover the entire acquisition volume 10 and to have three-dimensional information acquired substantially at the same time as each selected image, for example by operating continuously during the traversal of the hand part. In the presence of a three-dimensional acquisition system, the processing unit 106 then includes, for example: - a module for determining, in real time, three-dimensional information for the localization of said hand part within a coordinate system linked to the imaging system; and - a module for controlling the imaging system, and - a module for selecting acquired images. This preferential mode, particularly when outputting as a video stream from the side cameras, allows the selection of acquired images based on geometric position and image quality criteria.

[0094] Alternatively, a triggered mode is also possible and the last two listed modules can then be replaced by: - a module for determining the acquisition times ta, tb, te of the hand part passing in front of each lateral camera; and - a control module for the imaging system which then allows the acquisitions to be triggered at the acquisition times determined by the acquisition time determination module.

[0095] These modules can be hosted locally in electronic compartment 20 or on a remotely accessible server.

[0096] Figure 6 is an example of a schematic block diagram of an information processing unit 106 for implementing one or more embodiments of the invention. The information processing unit 106 typically comprises at least one calculator, computer, tablet, or other device enabling the execution of a computer program responsible for controlling the various stages of the process according to the invention. The information processing unit 106 includes a communication bus connected to: - a central processing unit 601, such as a microprocessor, denoted CPU; - a 602 transient memory, denoted RAM, to store the executable code of the method for implementing the invention, as well as registers adapted to record variables and parameters necessary for the implementation of the process according to embodiments of the invention; the memory capacity of the device can be supplemented by optional RAM memory connected to an expansion port, for example; - a non-transient memory 603, denoted FLASH, for storing computer programs and calibration data for the implementation of the embodiments of the invention; the stored computer programs include in particular a computer program comprising instructions adapted to the implementation of each of the steps of the process according to the invention when said program is executed on the processing unit 106, said FLASH memory 603 is then an example of a non-transient means of storing information, removable or not, partially or totally readable by a computer or a microprocessor comprising code instructions of the computer program for the execution of each of the steps of the process according to the invention; - A 604 network interface, denoted NET, is normally connected to a communication network over which digital data to be processed is transmitted or received. The 604 network interface can be a single network interface or composed of a set of different network interfaces (e.g., wired and wireless, or different types of wired or wireless interfaces). Data packets are sent over the network interface for transmission or are read from the network interface for reception under the control of the software application running in the 601 processor. - A 605 GUI user interface is used to receive input from a user or to display information to a user, including guidance information (voice and / or visual). - an I / O module 607 for receiving / sending data to / from external devices such as a hard drive, removable storage media, or others.

[0097] The executable code can be stored in non-volatile memory 603, for example, flash memory or read-only memory, or on removable digital media such as, for example, a disk. According to one embodiment, the executable code of the programs can be received via a communication network, through the network interface 604, in order to be stored in one of the storage means of the information processing unit 106, such as the FLASH memory 603, before being executed.

[0098] The central processing unit 601 is adapted to command and direct the execution of instructions or portions of software code of the program or programs according to one of the embodiments of the invention, instructions which are stored in one of the aforementioned storage means, such as the FLASH memory 603. After power-up, the CPU 601 is capable of executing instructions from the RAM memory Transient 602, relating to a software application. Such software, when executed by the processor 601, causes the execution of the described process.

[0099] The information processing unit 106, as illustrated, is local but can also be distributed and comprise multiple processing subunits, including physically remote subunits (outside the terminal) communicating with each other via the network interface. Similarly, part of the memory can be physically remote, hosted, for example, on servers. For instance, the module(s) for determining and / or the module for reconstructing a merged image may not be hosted locally but in a physically remote processing unit, so as to share existing computing power or to make the terminal even more compact.

[0100] Figure 7 illustrates the method E for non-contact acquisition of dermatoglyph images in an embodiment of the invention described in connection with the acquisition device 1 illustrated in Figures 3 and 4, without a central camera. The method E for non-contact acquisition of dermatoglyph images of a part of a user's hand (here a single finger for clarity) is described here, said part of the hand passing through an acquisition volume in a translational movement in a lateral direction parallel to the acquisition plane P, and advantageously comprising the following steps: - initialization El of the acquisition device 1 when it is switched on and triggering, via the control module, of the continuous acquisition (video stream) of the cameras CL1, CL2 of the imaging system; - lighting E2 of the acquisition volume 10, by means of lighting 60, so as to illuminate the part of the user's hand, for the qualitative acquisition of two-dimensional images; - acquisitions E3-1, E3-2, E3-3 (non-limiting number, at least two) of dermatoglyph images of the hand part acquired by the lateral cameras CL1,CL2, whose optical axes A21, A22 intersect upstream of the acquisition plane, at a plurality of instants ta, tb, te different from the translational movement of the hand part, crossing an acquisition volume 10 according to a translational movement in a lateral direction parallel to the acquisition plane P. Each of these acquisitions is not made by the same camera CL1,CL2 of the imaging system, so that all the acquisitions are not made according to the same angle of view but on the contrary these acquisitions are made along different optical axes A21, A22, intersecting upstream of the acquisition plane P, that is to say between the optical center proper to each optical axis and the acquisition plane P.The acquired images therefore form a series of images comprising a plurality of views of the hand part corresponding to different positions of said hand part in the acquisition volume 10. Among these acquisitions at least two acquired images image. a common area of ​​the hand part, said at least two images, called selected in a so-called non-triggered embodiment or triggered in a so-called triggered embodiment, being acquired at at least two different instants of the movement and along the different optical axes A21, A22; - E4 reconstruction of a merged image of a surface to be represented of said hand part from the selected / triggered acquired images, including in particular: - an extraction of characteristic points specific to said part of the hand in each acquired selected / triggered image; - a matching of common characteristic points between each pair of selected / triggered acquired images; - an assembly (also called stitching in English) of all or part of the selected / triggered acquired images to join parts of said images according to the correspondences determined in the previous step; - a determination, from the assembly, of an edge-to-edge representation of the dermatoglyph of the hand part.

[0101] The method E makes it possible to obtain different images of the hand part but in which there is an overlapping area of ​​the hand part so that the images complement each other for the purpose of fusion, each hand part being acquired along at least two different intersecting optical axes upstream of the acquisition plane P. This makes it possible to have at least two different images of the same hand part, in particular of each finger and more precisely of each dermatoglyph specific to each finger, and in each a common area of ​​the hand part is imaged so that the images can complement each other for the purpose of fusion, that is to say that said common imaged area is included in said at least two images.

[0102] Considering a common area per dermatoglyph of the hand part, each finger having a unique dermatoglyph, the common area is specific to each finger dermatoglyph, meaning there are as many common areas as there are fingers, and the same applies to the palm. For a finger, from the plurality of acquired images, at least two images of each finger are selected, acquired at different times from different viewing angles and imaging a common area of ​​the finger dermatoglyph. In other words, each finger is imaged at least at two distinct times during its passage through the acquisition volume, that is, in at least two different positions relative to the reference frame of the acquisition device 1, forming at least two distinct views of the surface to be represented of the finger, and this is done for each finger. In the case of the palm, from the plurality of acquired images, at least two images of the palm are selected, acquired at different angles. different moments from different angles of view and illustrating a common area of ​​the dermatoglyph of the palm.

[0103] The presentation of the selected images may constitute an edge-to-edge representation, or the selected images may then be used to obtain an edge-to-edge representation in the form of several two-dimensional unfoldings of the surface to be represented: for example, as many unfoldings as there are fingers, or for a palm, a single edge-to-edge unfolding; or it may take the form of a three-dimensional visualization of the surfaces to be represented, grouped into a single three-dimensional representation or split into multiple three-dimensional representations, for example, one per finger. When the part of the hand comprises at least two fingers, preferably four fingers, or a palm, the method E includes a step of segmenting the acquired selected images by finger.

[0104] Advantageously, a step of displaying the edge-to-edge representation of the hand part can be carried out on the device 1 if the latter has a screen, or can be implemented on a remote terminal.

[0105] Furthermore, by way of example and without limitation, the method includes, shortly after initialization or continuously throughout the entire process, a user guidance step prompting the user to perform a translational movement of said hand part within the acquisition volume 10, for example by means of instructions issued by the user interface 605, including voice guidance and / or a display, for example by means of a screen (displaying, for example, a translational movement of a hand through the acquisition volume), this display being able to include holographic elements. For example, during the initialization step 11, a first instruction is given to the user prompting them, for example, to pass their hand M with fingers D through the median plane of the acquisition volume 10 in a lateral translational movement, spreading the fingers.Indeed, a hand position with the four fingers spread apart makes the edges of each finger more visible to the imaging system. Guidance can also be achieved using light-emitting diodes (LEDs) embedded in the front of the frame and / or on the wall at the back of the user compartment and / or on the hood 15 along a line parallel to the acquisition plane P, so as to visualize the expected translational movement, and in particular its speed, by the successive flashing of the aligned LEDs. Alternatively, the user can be guided by an operator.

[0106] In the illustrated implementation of process E, the initialization step El triggers the illumination E2 of the acquisition volume 10 so as to illuminate the user's fingers D, and this illumination is preferably maintained until the end of process P. Alternatively, the illumination could be triggered by a proximity sensor so as not to illuminate when not in use. Alternatively, Lighting is only provided during image acquisition, itself triggered, for example, only in the presence of a user's hand, detected by means of a presence sensor, which helps to limit energy consumption.

[0107] The acquisition frequency of the lateral cameras CL1, CL2 is defined so as to acquire more than five images during translation to ensure that among the images each finger appears in at least one pair of images.

[0108] The non-triggered embodiment is preferred, involving the selection of acquired images from among a larger pool of acquired images, including a video stream. The selection is based primarily on geometric criteria (presence of the complete phalanx in the acquired image) and quality criteria (absence of aberration). This implementation allows for a posteriori selection, i.e., after the hand has passed, thus smoothing the CPU load. When a three-dimensional information acquisition system is present, the acquisition times ta, tb, te to be selected are advantageously determined based on the trajectory defined in a coordinate system linked to the finger (or palm) imaging system over time, based on the acquired three-dimensional information.Indeed, when a three-dimensional information acquisition system is present, it is preferentially activated during the initialization step (E1), and only stops after the hand portion has been scanned. Furthermore, the reconstruction step (E4) of a merged image is then facilitated by taking three-dimensional information into account.

[0109] Alternatively, in the triggered embodiment, acquisitions are controlled individually or in bursts and do not require selection among the images, even though selection based on quality factors is advantageous, for example, when implementing a triggered burst mode. Their control can be correlated with the guidance, considering that the translational movement represented by the latter is that performed by the user, and in particular synchronized with the illumination of the guide LEDs, which can be different colors for each finger in the case of multi-finger acquisition. Furthermore, in the presence of a three-dimensional information acquisition system, the acquisition control can be triggered based on the estimated trajectory of the fingers (or palm) using the acquired three-dimensional information.Indeed, in the presence of a three-dimensional information acquisition system, the latter is preferentially activated during the initialization step 11, and only stops after the hand part has passed through the acquisition volume 10. The location of said hand part, for example each finger, in a coordinate system linked to the imaging system is then determined in real time, based on the continuously received three-dimensional information. Advantageously, the acquisition times ta, tb, te are determined according to the geometric arrangement of the imaging system and as soon as a time . The acquisition point is determined—that is, the current position of the hand part at time t (or the estimated position of the hand part at a nearby time t) is within the acquisition field of a side camera—the imaging system is controlled to trigger the acquisition by said side camera at that time t (or approximately at that time t), and so on for subsequent times, and for each finger (or palm part). In burst mode, several acquisitions can thus be triggered approximately at time t for the same finger with the same camera, so that it is possible, for example, to select the highest quality image(s). The E4 reconstruction step of a merged image is then facilitated by taking into account the three-dimensional information.

[0110] The acquisitions E3-1, E3-2, E3-3 by the side cameras (resulting from a selection or a triggering) respectively at times ta, tb, te are for example recorded in the transient memory 602 and used asynchronously by the central unit 601 of the information processing unit 106, as would be the three-dimensional information in the presence of a three-dimensional information acquisition system.

[0111] The E4 reconstruction step of a merged image is performed by the 601 central processing unit. Part of the E4 reconstruction step can be synchronous with the reception of each image, in particular the extraction of feature points. The instructions adapted for implementing each of the matching, stitching, and edge-to-edge representation of the hand portion steps can be hosted locally in RAM or on a remote server.

[0112] Figure 8 illustrates the views acquired by cameras CL1, CL2, and CC of a device 1 according to an embodiment of the invention related to Figure 5, i.e., in the presence of a central camera and in non-triggered mode. In the case illustrated here, for the sake of clarity, we will focus on the middle finger and describe the steps of the process E in this embodiment related to the acquisition of the image of the middle finger D as sub-elements of the four fingers of the hand M, the same approach being applied to the four fingers, the process advantageously incorporating a segmentation step per finger.

[0113] At the top appears the time axis t, with the instants ta, tb, te of the acquisitions E3-1, E3-2 and E3-3.

[0114] The first line of images shows a chronological series, according to time t, of eight images (not exactly consecutive here because not all are shown) acquired by the central camera CC during the translational movement, or scrolling, of a user's right hand, from right to left. In other words, these eight images acquired at a plurality of instants form a series of images comprising a plurality of views of the middle finger (and even here of all the fingers of the hand) corresponding to different positions of the finger(s) in the free space of the acquisition volume, acquired by the central camera CC during the scrolling of the middle finger in the acquisition volume 10.

[0115] The second line of images shows the images acquired by the lateral camera CL2, imaging the left side of finger D at times close to the time of acquisition of the second image (as displayed) by the central camera. In other words, these three images acquired at a plurality of times form a series of images comprising a plurality of views of the middle finger corresponding to different positions of the middle finger in the free space of the acquisition volume, acquired by the central camera CC during the movement of the middle finger within the acquisition volume 10.

[0116] The third row of images shows the images acquired by the lateral camera CL1, imaging the right side of finger D at times close to the time of acquisition of the sixth image (as displayed) by the central camera. In other words, these three images acquired at a plurality of times form a series of images comprising a plurality of views of the middle finger corresponding to different positions of the middle finger in the free space of the acquisition volume, acquired by the central camera CC during the movement of the middle finger within the acquisition volume 10.

[0117] The vertical dotted lines on the images acquired by the central camera CC or lateral cameras CL1, CL2 serve only to designate the middle finger D on these images for better reading of the figure.

[0118] On the fourth line of images, the middle finger D is schematically represented in the images acquired by the imaging system during the three acquisitions E3-1, E3-2, E3-3 at times ta,tb,tc to better illustrate the position and orientation of the middle finger D relative to the camera which acquired the image of it, in particular relative to the center of the dermatoglyph of the middle finger, represented in the form of a spiral, the dotted area representing the surface of the middle finger D which was acquired in addition to the frontal image acquired E3-2, so that the total surface to be represented extends well edge to edge of the dermatoglyph.

[0119] It should be noted that not all four fingers of the hand appear on all the images acquired by the central camera CC and even less on all the images (here, on none) acquired by the side cameras CL1, CL2.

[0120] In the embodiment shown in this figure, the central camera CC continuously acquires (video stream) images covering the width of the 14 cm acquisition volume, and a selection is made from these images, notably according to quality criteria, to choose at least two images of the middle finger (corresponding to at least two instants: one at the beginning of the middle finger's passage through the acquisition surface to view the middle finger through a lateral camera and the other at the end of the crossing of the acquisition surface by the middle finger to see the middle finger by the other lateral camera), or preferably (as here) at least three images of the middle finger corresponding to at least three instants, namely here: a first instant (~ta) at the beginning of the crossing of the acquisition surface by the middle finger to see the middle finger by a lateral camera, a second instant tb at mid-crossing to see the middle finger from the front by the central camera and a last instant (~tc) at the end of the crossing of the acquisition surface by the middle finger to see the middle finger by the other lateral camera.

[0121] Once these times have been determined, the images acquired by the relevant side camera at times identical (particularly if the cameras are synchronized) and / or close (if the cameras are not synchronized) to the times of acquisition by the central camera CC of the second and sixth images (as displayed) are extracted from RAM. In the example, three images acquired by the side camera CL2, imaging the left side of the middle finger D at times close to time ta, are shown. It can be seen that the middle finger is not visible in the first image, and the second of these images is selected here, its acquisition corresponding to acquisition step E3-1, here by the side camera CL2 at time ta. Acquisition step E3-2 corresponds here to the acquisition by the central camera CC at time tb.Similar to step E3-1, three images acquired by the lateral camera CL1, imaging the right side of the middle finger D at times close to time te, are shown. Among these images, the second is selected here, its acquisition corresponding to acquisition step E3-3, here by the lateral camera CL1 at time te. A fortiori, by selecting such images, the three images acquired at the plurality of times ta, tb, te in steps E3-1, E3-2, E3-3 form a series of images comprising a plurality of views of the middle finger corresponding to different positions of the middle finger D in the free space of the acquisition volume 10, acquired here by each of the three cameras CL1, CL2, CC of the imaging system during the movement of the middle finger within the acquisition volume 10.

[0122] This embodiment allows that for each finger D an image is acquired by each camera: central and lateral CL1,CL2,CC, each at a different position of the finger, and used to perform the fusion.

[0123] Consequently, a wide-field image is acquired by the central camera at approximately the same time as the acquisition time specific to each of the selected images acquired by one of the lateral cameras CL1, CL2. This allows its wider-field image to be used to assist image fusion, particularly in the case of multi-finger applications, to establish the correspondence between the fingers and the acquired images, so as to know which lateral camera images which finger, thus facilitating the finger segmentation step. Similarly, for multi-finger acquisition, at a minimum, the phalanx of a finger, preferably The phalanges of several fingers are seen in their entirety in the selected image acquired at the acquisition time specific to each camera in the imaging system. Advantageously, the selected images include views of the entire finger (phalange, phalanx, and phalanx).

[0124] As shown in the fourth line, the three images acquired in steps E3-1, E3-2, and E3-3 thus share (here in all images) imaged points of the dermatoglyph of the middle finger. This common area ensures partial overlap between these images, allowing for reconstruction E4 of the fused image of the surface to be represented of the hand part, in this case the middle finger, and then similarly for the four fingers, from the three acquired images. This reconstruction is achieved, in particular, by matching the common characteristic points between each pair of views in the plurality of views, and then assembling the three images according to the determined matchings. It should be noted that the number of three images is not limiting; preferably, five images are used.

[0125] Since the reconstruction step E4 of a fused image of a surface to be represented of said part of hand from the acquired images is optional, the presentation of the selected images E3-1, E3-2, E3-3 can directly constitute an edge-to-edge representation (without stitching).

[0126] An additional image of the hand part can also be acquired by at least one of the cameras, lateral or central, during an insertion movement of the hand perpendicular to the translational movement, in particular prior to the translational movement, which makes it possible to obtain for example an image of 4 fingers and / or of the palm and this additional image can also be used in the reconstruction of the fused image, or simply serve to confirm the correspondence with the acquired curled fingers.

[0127] Alternatively, rather than selecting camera images from a video stream after the fact, it is also possible to control the side cameras in a more targeted manner, without the side cameras producing a continuous video stream. Instead, the side cameras CL1 and CL2 are controlled, notably by means of the processing unit 106, so as to trigger their acquisition—either one-off or in bursts over a very short period relative to the time it takes the hand to traverse the acquisition surface—based on the content (analyzed in real time) of the images continuously acquired (video stream) by the central camera or by a three-dimensional information acquisition system for the acquisition volume. This variant consumes fewer resources.

[0128] Naturally, to satisfy specific needs, a person competent in the field of the invention may apply modifications to the preceding description.

[0129] Although the present invention has been described above with reference to specific embodiments, the present invention is not limited to specific embodiments, and modifications which fall within the scope of the present invention will be obvious to a person versed in the art.

Claims

Demands

1. A non-contact dermatoglyphic imaging device (1) for at least a portion of a user's hand (M, D), said acquisition device comprising: - an acquisition volume (10) extending on either side of an acquisition plane (P) and constituting a free space area forming a passage for said hand portion (M, D) moving in translation along a lateral direction parallel to the acquisition plane (P), and - an electronic compartment (20), said electronic compartment comprising a fixed imaging system having at least two side cameras (CL1, CL2), each having its own acquisition field partially covering the acquisition volume, said acquisition fields being distinct with an overlap, defined as an intersection surface on a union, in the acquisition plane of 0 to a maximum of 50%, said side cameras (CL1, CL2) having optical axes (A21,A22) secants intersecting upstream (I) of the acquisition plane (P) characterized in that: each of said lateral cameras (CL1, CL2) is configured to acquire an image of said part of the hand (M,D), each of the lateral cameras acquiring said image at a different instant (ta, te) of the lateral translation movement of said part of the hand (M,D) in the acquisition volume (10), each instant (ta, te) being specific to each lateral camera (CL1, CL2), said images acquired at said instants (ta, te) forming a series of images comprising a plurality of views of said part of the hand (M,D) corresponding to different positions of said part of the hand (M,D) in the free space of the acquisition volume (10) and imaging a common area of ​​said part of the hand (M,D).

2. Device (1) according to claim 1, wherein at least one hand part (M,D) comprises at least two fingers (D), preferably four fingers (D), or a palm.

3. Device (1) according to any one of claims 1 to 2, wherein the imaging system comprises a central camera (CC) whose acquisition field covers all or part of the acquisition fields of the side cameras (CL1,CL2) in the acquisition plane (P), and preferably covering the entire acquisition volume (10), said central camera (CC) being in particular positioned so that an intersection at the level of the acquisition plane (P) of a projection of an optical axis (A2C) of said central camera (CC) in a plane defined by the optical axes (A21, A22) of two of the side cameras (CL1,CL2) is located between the intersections of the optical axes (A21, A22) of said two side cameras with the acquisition plane (P).

4. Device (1) according to any one of claims 1 to 3, wherein a processing unit (106) is connected to the imaging system, the processing unit (106) being configured to fuse the acquired images imaging the common area.

5. Device (1) according to any one of claims 1 to 4, wherein the angular aperture (Q1, Q2) of the acquisition fields of each of the lateral cameras (CL1, CL2) is less than 25°, in particular less than 15° and preferably substantially equal to 10°

6. d lu . Device (1) according to any one of claims 1 to 5, wherein in a plane common to the optical axes (A21,A22) of each of the side cameras (CL1,CL2), the optical axis of at least one of the side cameras (CL1,CL2) is inclined with respect to a projection of a normal to the acquisition plane in said common plane between 10° and 60°, preferably between 25° and 45°.

7. Device (1) according to any one of claims 1 to 6, wherein two of said side cameras (CL1,CL2) are positioned symmetrically with respect to a cutting plane of the device (1).

8. Device (1) according to any one of claims 1 to 7, wherein the electronic compartment (20) includes a lighting means (60) capable of illuminating the acquisition volume (10), said lighting means (60) comprising at least two light sources, each of which has a mean diffusion axis substantially collinear with the optical axis of each camera (CL1,CL2,CC).

9. Device (1) according to any one of claims 1 to 7, wherein the electronic compartment (20) comprises a lighting means (60) capable of illuminating the acquisition volume, said lighting means (60) comprising a plurality of light sources, arranged annularly around each camera (CL1,CL2,CC) or between the cameras (CL1,CL2,CC).

10. A method (E) for non-contact acquisition of dermatoglyphic images of at least a part of a user's hand (M,D), said hand part (M,D) traversing an acquisition volume (10) in a translational movement in a lateral direction parallel to an acquisition plane (P), said method (E) comprising the acquisition (E3-1, E3-2, E3-3) of images of said hand part (M,D) at a plurality of times (ta, tb, te) different from the translational movement of said hand part (M,D), said images acquired at the different times (ta, tb, te) forming a series of images comprising a plurality of views of said hand part corresponding to different positions of said hand part (M,D) in the acquisition volume (10) at a plurality of times of the movement and at least two of said images are acquired along an optical axis (A21, A22) different, the said optical axes (A21,A22) intersecting (I) upstream of the acquisition plane (P) and said at least two of said images imaging a common area of ​​said part of hand (M,D).

11. Method (E) according to the preceding claim, said method being implemented by an acquisition device (1) according to any one of claims 1 to 9.

12. Method (E) according to any one of claims 10 to 11, wherein said at least one hand part (M,D) comprises at least two fingers (D), preferably four fingers, or a palm, the common area being specific respectively to each finger or palm.

13. Method (E) according to claim 12, said method comprising a finger segmentation step (D) of the acquired images.

14. Method (E) according to any one of claims 10 to 13, said method comprising a reconstruction step (E4) of a fused image of a surface to be represented of said part of hand from the acquired images.

15. Method (E) according to claim 14, the reconstruction step (E4) of the image merged from the acquired images comprising: - an extraction of characteristic points specific to said hand part (M,D) in each image of the plurality of views; - a matching of the common characteristic points between each pair of views of the plurality of views; - an assembly of the plurality of dermatoglyphic images to join parts of said images according to the correspondences determined in the previous step; - a determination, from the plurality of assembled dermatoglyphic images, of an edge-to-edge representation of the dermatoglyphic of the hand part.

Images