Ultrasonic transducer biometric identification device with fingerprint and microvasculature detection

Abstract

The invention relates to a biometric identification device comprising an ultrasonic transducer circuit, a control circuit and a data processing circuit configured to implement: - acquiring a surface image of an element arranged against a capture surface, and a series of volume images of the element from which an overall volume image (114) of a microvasculature of the element is determined; - determining features of a fingerprint of the element from the surface image, and features of the microvasculature of the element from the overall volume image (114); - comparing the previously determined features with expected theoretical features in order to identify or not identify the element.

Description

DESCRIPTION TITLE: Biometric identification device with ultrasonic transducer for fingerprint and microvascular detection The present application claims priority from French patent application number 23 / 10033, filed on September 22, 2023, entitled "Biometric identification device with ultrasonic transducer for fingerprint and microvascular detection", which is incorporated by reference to the fullest extent permitted by law. technical field

[0001] This description relates to the field of biometric identification devices based on ultrasonic transduction capture and detection. Previous technique

[0002] A biometric identification device verifies or determines the identity of a device user based on a measurement of one or more of the user's biometric characteristics, such as a fingerprint, facial shape, iris pattern, retinal pattern, etc. The various techniques for measuring these characteristics each have advantages and disadvantages depending on several criteria, including: error rate (more or less significant), ease or difficulty of theft of the template against which the measurements are compared, ability to detect forgeries, proof of life (or lack thereof), ease and comfort of use, required sensor size, and energy consumption.

[0003] It is known to create a biometric identification device, an anti-authentication device (“anti- spoofing » in English) and liveness detection using a fingerprint sensor performing optical or capacitive measurements to verify if the captured fingerprint corresponds to the expected one (identity verification phase), then implementing a liveness detection and anti-spoofing algorithm via image processing performed from the captured fingerprint or using an infrared system allowing, for example, to detect if the finger placed on the device during fingerprint capture is a fake finger. Summary of the invention

[0004] There is a need to propose a biometric identification system that does not present at least some of the disadvantages of existing solutions.

[0005] One embodiment overcomes all or part of the drawbacks of known solutions and proposes a biometric identification device comprising at least one ultrasonic transducer circuit, a control circuit and a data processing circuit, configured to implement at least the following steps:

[0006] - acquisition of at least one surface image of at least one element intended to be placed against a capture surface of the device, and of at least one series of volume images of the element from which an overall volume image of a microvasculature of the element is determined;

[0007] - determination of characteristics of an imprint of the element from the surface image, and of characteristics of the microvasculature of the element from the overall volume image;

[0008] - comparison of the previously determined characteristics, i.e., those determined in the previous step, with expected theoretical characteristics, such as characteristics calculated during a preliminary user enrollment stage, to identify or not the 'element'.

[0009] According to a particular embodiment, the data processing circuit is configured to determine the overall volume image by implementing a spatio-temporal filtering process of the series of volume images of the element.

[0010] According to a particular embodiment, the ultrasonic transducer circuit is configured to acquire the series of volume images of the element by pulsed Doppler imaging.

[0011] According to a particular embodiment, the data processing circuit is configured to implement, between the step of acquiring the surface image and the series of volume images and the step of determining the characteristics of the footprint and microvascularity of the element, a step of aligning the surface image and the overall volume image with respect to each other.

[0012] According to a particular embodiment, the data processing circuit is configured such that the step of determining the characteristics of the footprint and microvascular structure of the element includes at least:

[0013] - determination of minutiae of the element's imprint from the surface image, and of points of interest of the element's microvasculature from the overall volume image;

[0014] - calculation of the position of at least one minutiae in the surface image and / or at least one point of interest in the overall volume image, and of geometric relationships between said at least one minutiae and / or said at least one point of interest and at least some of the other minutiae and points of interest, taking into account, for the points of interest, the depths at which the points of interest are located in the overall volume image;

[0015] and such that the comparison step involves a comparison of the previously calculated position and geometric relationships with an expected theoretical position and geometric relationships to identify or not the element.

[0016] According to a particular embodiment, the points of interest correspond to bifurcations of micro-blood vessels and / or to points of strong curvature of micro-blood vessels.

[0017] According to a particular embodiment, the data processing circuit is configured to implement, between the stage of determining minutiae and points of interest and the calculation stage, a step of combining, in the form of a point cloud, the minutiae and points of interest determined.

[0018] According to a particular embodiment, the data processing circuit is configured such that, during the calculation step, the other minutiae and points of interest with which the geometric relationships are calculated correspond to the minutiae included in a region of the surface image delimited by a planar geometric figure and to the points of interest included in a region of the global volume image included in a volume formed from the planar geometric figure which is extended in a dimension of the global volume image perpendicular to the surface image.

[0019] According to a particular embodiment, the plane geometric figure corresponds to a polygon, a circle or an ellipse.

[0020] According to a particular embodiment, the data processing circuit is configured such that the step of calculation is repeated for each minutiae and each point of interest determined, and the comparison step is performed for the position and geometric relationships calculated for each minutiae and each point of interest.

[0021] According to a particular embodiment, the data processing circuit is configured such that the geometric relationships calculated between a first minutia or point of interest and a second minutia or point of interest include the distance between the first minutia or point of interest and the second minutia or point of interest, an orientation of the first minutia or point of interest relative to the second minutia or point of interest, and an orientation of the second minutia or point of interest relative to the first minutia or point of interest.

[0022] According to a particular embodiment, the data processing circuit is configured such that, for a point of interest corresponding to a bifurcation or a point of strong curvature of micro-blood vessels, the calculated geometric relationships include the orientations of each of the micro-blood vessels with respect to the other minutiae and points of interest.

[0023] According to a particular embodiment, the data processing circuit is configured as follows:

[0024] - the determination of the characteristics of the footprint and microvasculature of the element involves the extraction of at least one vector of features from the surface image and the overall volume image by the implementation of a deep learning algorithm;

[0025] - the comparison of the previously determined characteristics with expected theoretical characteristics to identify or not the element involves a comparison, by the deep learning algorithm, of the vector of characteristics previously determined with at least one expected theoretical characteristic vector.

[0026] According to a particular embodiment, the data processing circuit is configured such that the comparison step includes a score calculation whose value depends on the correlations between the determined characteristics and the theoretical characteristics expected to identify or not the element.

[0027] According to a particular embodiment, the data processing circuit is configured such that, during the comparison step, a score is calculated for each minutiae and each point of interest determined. Brief description of the drawings

[0028] These features and advantages, as well as others, will be described in detail in the following description of particular embodiments, given by way of non-limiting example, in relation to the attached figures, among which:

[0029] - Figure 1 and Figure 2 schematically represent a biometric identification device according to a particular embodiment;

[0030] - Figure 3 represents, in the form of a schematic diagram, the steps implemented by a biometric identification device according to a particular embodiment;

[0031] - Figure 4 schematically illustrates minutiae identified in a surface image captured by a biometric identification device according to a particular embodiment;

[0032] - Figure 5 schematically represents a region delimiting the minutiae and points of interest considered during an identification performed by a device biometric identification according to a particular embodiment;

[0033] - Figure 6 schematically represents an example of calculation results obtained when the identity of the user of the biometric identification device is confirmed;

[0034] - Figure 7 schematically represents a first example of calculation results obtained when the identity of the user of the biometric identification device is not confirmed;

[0035] - Figure 8 schematically represents a second example of calculation results obtained when the identity of the user of the biometric identification device is not confirmed. Description of the implementation methods

[0036] The same elements have been designated by the same reference numerals in the different figures. In particular, structural and / or functional elements common to the different embodiments may have the same reference numerals and may have identical structural, dimensional and material properties.

[0037] For the sake of clarity, only the steps and elements necessary for understanding the described embodiments have been shown and are detailed. In particular, various components (ultrasonic transducer circuit, control circuit, processing circuit, etc.) and various steps implemented (processing of acquired images, determination of characteristics, details of calculations performed, etc.) are not detailed. Those skilled in the art will be able to implement these components in detail based on the functional description provided here.

[0038] Unless otherwise specified, when referring to two connected elements, this means directly connected without any intermediate elements other than conductors, and when referring to two coupled elements, this means that these two elements can be connected or linked through one or more other elements.

[0039] In the description that follows, when referring to absolute positional qualifiers, such as the terms "front", "back", "top", "bottom", "left", "right", etc., or relative positional qualifiers, such as the terms "above", "below", "superior", "inferior", etc., or to orientational qualifiers, such as the terms "horizontal", "vertical", etc., unless otherwise specified, reference is made to the orientation of the figures in a normal position of use.

[0040] Unless otherwise specified, the expressions "approximately", "roughly", "about", and "on the order of" mean within 10%, preferably within 5%.

[0041] A biometric identification device 100 according to a particular embodiment is described below in relation to figures 1 and 2.

[0042] In the example shown in Figure 1, the device 100 is configured to perform biometric identification using an element 101 of the device 100 user, this element 101 corresponding, for example, to one or more of the user's fingers. The element 101 here corresponds to an element, or part, of the body of a device 100 user. The device 100 includes a capture surface 102 on which this element 101 of the device 100 user is intended to be placed during identification.

[0043] The device 100 comprises at least one ultrasonic transducer circuit 104, a control circuit 106, and a data processing circuit 108. In the schematic example shown in Figure 2, the control circuit 106 can be electrically coupled to the ultrasonic transducer circuit 104 and thus be capable of transmitting signals to the ultrasonic transducer circuit 104, controlling the emission of ultrasound for the measurements to be performed, and also capable of receiving electrical measurement signals transmitted by the ultrasonic transducer circuit 104. The control circuit 106 can also be coupled to the data processing circuit 108 in order to transmit to the data processing circuit 108 the measurement results obtained from the ultrasonic transducer circuit 104.

[0044] The circuit 104 can include a plurality of ultrasonic transducers enabling the image acquisitions required for biometric identification. The ultrasonic transducers are, for example, arranged in a matrix or in another manner adapted to the captures to be performed. The number of ultrasonic transducers in the circuit 104 may depend on the dimensions of the capture surface 102.

[0045] In the described embodiment, the device 100 can be configured to acquire at least one surface image of the user's finger(s) placed on and in contact with the capture surface 102 of the device 100; that is, an acquisition of at least one image of at least one fingerprint formed by the ridges and valleys present on the skin surface of the finger(s) placed on the capture surface 102. In the device 100, the acquisition of a surface image of the user's finger(s) placed on and in contact with the capture surface 102 is based on the fact that the ultrasound emitted by the transducer circuit 104 reflect more strongly against the air present in the valleys of the fingerprint of the finger(s) than against the ridges of the print.

[0046] Furthermore, the device 100 can also be configured to acquire volume images of the finger(s) placed on and in contact with the capture surface 102 of the device 100, and from which it is possible to determine, by processing these volume images, an overall volume image of the microvasculature of this finger(s), i.e. an image capture of the micro-blood vessels of the finger(s) of the user of the device 100. The overall volume image obtained can thus be deduced from several volume images captured successively at non-zero depths (dimension along the Z-axis visible in Figure 1 and which is, for example, substantially perpendicular to the capture surface 102) of the finger(s) present on the capture surface 102.

[0047] The ultrasonic transducer circuit 104 can, for example, be configured to acquire volume images of the finger(s) placed on the capture surface 102 using pulsed-wave Doppler imaging. In such a configuration, the transducer circuit 104 is configured to emit a series of ultrasonic pulses. The responses obtained in this configuration do not correspond to a change in the frequency of the received wave, but rather to a pseudo-Doppler effect, that is, a change in the time interval between the received echoes relative to the time interval between the emitted waves. In the device 100, pulsed-wave Doppler imaging can therefore be used to determine the spatial location of microvascular movements in the acquired volume images and at from which a microvasculature map is determined in the global volume image.

[0048] The ultrasonic transducer circuit 104 may, for example, include CMUT-type transducers (in English, "Capacitive Micromachined Ultrasonic Transducer" or PMUT type ("Piezoelectric Micromachined Ultrasonic Transducer" or other type.

[0049] In the example embodiment described here, the control circuit 106 can be configured to provide the ultrasonic transducer circuit 104 with electrical excitation signals causing the ultrasonic transducer circuit 104 to emit ultrasonic waves, and to receive electrical response signals generated by the ultrasonic transducer circuit 104 under the effect of receiving ultrasonic waves reflected by the element 101 present on the capture surface 102.

[0050] In the described embodiment, the data processing circuit 108 can be configured to analyze and process the electrical response signals received by the control circuit 106 and sent by the ultrasonic transducer circuit 104. The processing circuit 108 may include, for example, a microprocessor coupled with a memory to perform the processing of the received data.

[0051] The control circuit 106 and the data processing circuit 108 can be configured to perform biometric identification by implementing the steps described below and represented in diagram form in Figure 3.

[0052] In a first step 202, the device 100 can perform a detection of an element present on the capture surface 102. If an element is indeed detected on the capture surface 102, a second step 204 can be implemented. If no element is detected, this detection step 202 can be repeated until an element is detected on the capture surface 102. The detection step 202 can be repeated continuously or at regular or irregular time intervals.

[0053] The device 100 can then be configured to acquire at least one surface image of the element 101 present on the capture surface 102 (step 204). For example, this configuration of the device 100 may include sending an acquisition command, or a sequence of instructions, to the control circuit 106 for acquiring the surface image of the element 101.

[0054] In a subsequent step 206, an acquisition of at least one surface image of the element 101 present on the capture surface 102 can then be implemented. This acquisition can include the emission of a series of ultrasonic signals in the form of pulses by the transducer circuit 104, then the reception of the echoes by the transducer circuit 104 and the processing of the responses obtained to obtain the surface image of the element 101, i.e. the image of the user's fingerprint(s).

[0055] The processing of the responses obtained to acquire the surface image of element 101 corresponds, for example, to signal processing of the "beamforming" or "microbeamforming" type, also called beamforming or spatial filtering. Alternatively, it is possible to transmit and receive signals from the transducer circuit 104 by multiplexing, to acquire a B-mode image focused on the surface of element 101. According to another variant, it is possible to perform a Orthogonal addressing of the transducer matrix of circuit 104 to emit a series of ultrasonic pulses in the form of focused waves, or plane waves, or to perform a complete matrix capture of the RCA-OPW type ("Row column addressing orthogonal plane wave imaging"). As an example, the document by V. Perrot et al., "So you think you can DAS? A viewpoint on delay-and-sum beamforming", Ultrasonics 111 (2021), 106309, arXiv: 2007.11960, describes an example of a DAS (Delay-and-sum) beamforming method that can be used to obtain the surface image of element 101.

[0056] The acquisition of the surface image of element 101 may also involve the implementation of other steps not detailed here: filtering of the response signal obtained, envelope detection, logarithmic compression, etc.

[0057] The acquired surface image can correspond to the capture of an imprint of element 101, for example, one or more fingerprints when element 101 corresponds to one or more fingers of the user. The data relating to the acquired image is stored, for example, in the memory of device 100 or in an external memory of device 100 connected to device 100 via a communication link.

[0058] After acquiring this surface image of element 101, the device 100 can be configured to perform successive acquisitions over time of several volume images of element 101 present on the capture surface 102 (step 208). For example, this configuration of device 100 may involve sending an acquisition command, or a sequence of instructions, to the control circuit 106 for acquiring volume images of element 101.

[0059] In a subsequent step 210, several successive volume images of element 101 are acquired present on capture surface 102, these volume images can then be processed to highlight displacements or movements related to blood flow in the microvasculature of element 101 and ultimately obtain a global volume image of the microvasculature of element 101. The processing applied to the volume images can involve spatiotemporal filtering, or "clutter filtering," using high-pass filtering in the so-called "slow-time" domain after signal demodulation. Alternatively, the estimation of movements related to blood flow in the microvasculature can be performed by correlating movements between successive volume images, or by SVD (singular value decomposition) and rejection of components that include tissue movement and noise. The document by Demené C et al."Spatiotemporal Clutter Filtering of Ultrafast Ultrasound Data Highly Increases Doppler and fUltrasound Sensitivity", IEEE Trans Med Imaging. 2015 Nov, 34 (11), 2271-85, describes an example of the implementation of such spatiotemporal filtering.

[0060] For example, volume image acquisition can be performed using plane-wave ultrasonic waves emitted by transducer circuit 104 at a number of angles. The processing of the response signals can then involve a beamforming phase followed by coherent summation of the images corresponding to each angle to obtain a B-mode volume image. This process can be repeated over time to obtain multiple volume images, allowing monitoring of blood flow in the microvasculature of element 101. Ultimately, after processing the volume images, a global volume image of the microvasculature of element 101 can be obtained. As an example, see the document by G. Montaldo et al., "Coherent plane-wave compounding for very high frame rate ultrasonography and transient elastography, in IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 56, no. 3, pp. 489-506, March 2009, describes an example of a DAS beamforming method with coherent summation of plane waves that can be implemented to obtain the global volume image of the microvasculature of element 101.

[0061] Thus, the microvasculature of element 101 is determined by first identifying, through the series of volume images acquired, the movements of tissues and arterial walls due to blood circulation allowing localization of the elements of the microvasculature.

[0062] The data relating to the overall volume image obtained can be stored in a memory of the device 100 or in an external memory of the device 100 for example connected to the device 100 by a communication link.

[0063] In the particular embodiment described, in a subsequent step 212, the previously acquired surface image and global volume image can be aligned with each other, for example by repositioning them according to a common coordinate system defined by the geometry of the transducer circuit 104.

[0064] In a subsequent step 214, using the previously acquired surface image, the data processing circuit 108 can determine or extract minutiae from the fingerprint of element 101 obtained from this image. This minutiae determination can be performed by one or more image processing algorithms not described in detail here but known to those skilled in the art. Various methods for extracting minutiae from a fingerprint image that can be used here are described in the paper by R. Bansal et al., “Minutiae Extraction from Fingerprint Images - a Review”, IJCSI International Journal of Computer Science Issues, Vol. 8, Issue 5, No 3, September 2011.

[0065] In a subsequent step 216, from the previously acquired global volume image, the data processing circuit 108 can determine or extract points of interest in the microvasculature of the finger(s) present on the capture surface 102. In the described embodiment, these microvasculature points of interest correspond to bifurcations of microvessels and / or points of strong curvature, or strong inflection, of microvessels. This determination of points of interest can be performed by one or more image processing algorithms, for example, similar to those implemented for the previous determination of the fingerprint minutiae. Alternatively, the algorithm implemented to determine the points of interest may include the following steps:

[0066] - segmentation of the global volume image, for example either by features of a Hessian matrix in each voxel of the global volume image, or by the use of a convolutional neural network or other;

[0067] - skeletonization ("skeletonization" or "thinning" in English) of the segmented image, with the aim of reducing the thickness of elements (for example blood vessels) present in the image to a voxel;

[0068] - detection of bifurcations and / or strong inflections;

[0069] - characterization of the orientation of each point of interest, for example by principal component analysis or DCA (“Principal component Analysis” in English).

[0070] According to another variant, the algorithm implemented to determine the points of interest may include the use of a neural network to directly detect points of interest as well as their type (bifurcation or strong curvature) and their orientation.

[0071] Alternatively, it is possible to acquire the surface image after the volume images have been acquired. In addition, it is also possible for the determination of minutiae in the surface image to be carried out before or during the acquisition of the volume images, or between the acquisition of the volume images and the processing of these images.

[0072] In the particular embodiment described, during a subsequent step 218, a combination step, in the form of a point cloud, of the previously determined minutiae and points of interest is implemented. This combination step may involve placing the imprint minutiae and the microvascular points of interest in the same coordinate system linked to the geometry of the transducer circuit 104. Thus, the imprint minutiae can be placed in an (X, Y) plane at a shallow depth or zero depth (Z = 0), and the microvascular points of interest can be positioned in a volume located below that in which the minutiae are located. This combination of data from previous image acquisitions is, for example, implemented by the data processing circuit 108.

[0073] The data processing circuit 108 can then calculate the position of at least one minutiae in the surface image and / or at least one point of interest in the overall volume image, and calculate geometric relationships between said at least one minutiae and / or said at least one point of interest and at least some of the other minutiae and points of interest, taking into account, for the points of interest, the depths (dimension along the Z-axis of the figure 1) which are the points of interest in the global volume image. In a particular embodiment, the data processing circuit 108 is configured such that, during this calculation step, the other minutiae and points of interest with which the geometric relations are calculated correspond to the minutiae included in a region of the surface image delimited by a planar geometric figure (for example a polygon, a circle or an ellipse) and to the points of interest included in a region of the global volume image delimited by the volume from this planar geometric figure and extended in the direction of the depth (parallel to the Z axis) of the global volume image.

[0074] An example of such a calculation is described below in relation to figures 4 and 5.

[0075] Figure 4 shows an example of a surface image acquired by device 100 and is designated by reference 110. This surface image 110 corresponds here to the fingerprint of a finger present on the capture surface 102 of device 100. The minutiae determined in this surface image 110 are represented by points and designated by references 112, 113.

[0076] An example of a global volume image acquired by the device 100 is shown in Figure 5 and designated by reference 114. In Figure 5, this image corresponds to the result obtained by maximum intensity projection (or MIP for "Maximum Intensity Projection") along the Z axis of the microvasculature of the finger present on the capture surface 102 of the device 100. The points of interest determined in this image are represented by points designated by reference 116.

[0077] In the example described here, one of the minutiae or one of the points of interest 116 can then be chosen, and then a region The surface image 110 and a region of the overall volume image 114, superimposed on one another and including this minutiae or point of interest 116, are delimited, on one side by a planar geometric figure for the surface image 110, and by a volume formed from the planar geometric figure for the overall volume image 114. In the example of Figures 4 and 5, one of the minutiae, designated by the reference 113, is chosen, and the planar geometric figure is defined as a circle whose center corresponds to the minutiae 113. The calculation step is implemented by considering the other minutiae 112 included in the geometric figure and the points of interest 116 included in a volume 120 formed by projecting this geometric figure along the Z-axis, that is, in the direction of the thickness of the element 101.Figure 5 schematically represents the volume 120 obtained and delimiting the region of the finger(s) in which the minutiae 112, 113 and the points of interest 116 are considered for this calculation. In the example of Figure 5, this volume 120 forms a cylinder extending through at least part of the thickness of the finger(s) present on the capture surface 102.

[0078] Thus, minutiae 112, 113 and points of interest 116 are projected into a common 3D space from which the following calculation steps are implemented.

[0079] The position of minutiae 113 in surface image 110 can be calculated, and geometric relationships between this minutiae 113 and the other minutiae 112 and the points of interest 116 included in volume 120 can also be calculated. In the example described, these geometric relationships may include:

[0080] - the distance between minutiae 113 and each of the other minutiae 112 and points of interest 116 included in volume 120, and

[0081] - an orientation of the minutiae 113 in relation to the other minutiae 112 and points of interest 116 included in volume 120, and

[0082] - an orientation of the other minutiae 112 and points of interest 116 included in volume 120 in relation to minutia 113.

[0083] This calculation can be implemented by considering all the minutiae 112 and points of interest 116 present in a volume whose section corresponds to the chosen geometric figure and which extends into at least part of the thickness of the finger(s) present on the capture surface 102.

[0084] In one particular embodiment, the data processing circuit 108 is configured such that the calculation step described above is repeated for each previously determined minutia 112 and point of interest 116. When this calculation is implemented for a point of interest 116 corresponding to a bifurcation or sharp bend in microvessels, the calculated geometric relationships may include the orientations of each of the microvessels relative to the other minutiae 112 and points of interest 116.

[0085] For the calculations described above, the positions of minutiae 112, 113 and points of interest 116 as well as the geometric relationships between them are calculated in Cartesian or spherical coordinates.

[0086] The implementation of this calculation step(s) corresponds to step 220 of the diagram in Figure 3.

[0087] The position(s) of minutiae 112, 113 and of points of interest 116 and the previously calculated geometric relationships are then compared with one or more previously calculated positions and geometric relationships, for example during a preliminary user enrollment step, to confirm or not the identity of the user of device 100 (step 222). This comparison can be performed for the positions and geometric relationships calculated for each minutia 112 and each point of interest 116.

[0088] Figure 6 schematically illustrates an example in which the calculated positions and geometric relationships correspond to those expected, thus confirming that the identity of the user of device 100 is indeed the expected one. In this figure, reference 122 designates the expected positions and geometric relationships between minutiae 124 and points of interest 126, which are similar to those calculated for minutiae 112, 113 and points of interest 116 of the finger present on the capture surface 102 of device 100.

[0089] Figure 7 schematically illustrates another example in which the calculated positions and geometric relationships do not correspond to those expected to confirm the identity of the user of device 100. In this example, the determined minutiae 112, 113, the positions of these minutiae, and the geometric relationships between them do correspond to those expected, but no point of interest 116 is determined (and therefore no geometric relationship with points of interest is calculated). Such an example corresponds, for instance, to the case where a fake finger is placed on the capture surface 102 of device 100. In such a case, device 100 may consider that the identity of the user of device 100 is not confirmed.

[0090] Figure 8 schematically illustrates another example in which the calculated positions and geometric relationships do not correspond to those expected to confirm the identity of the user of device 100. In In this example, the minutiae 112, 113 and the points of interest 116 determined, their positions and the geometric relationships between them do not correspond to those expected. Such an example corresponds, for instance, to the case where the finger placed on the capture surface 102 of the device 100 is a real finger but is not that of the person intended to be identified. In such a case, the device 100 may consider that the identity of the user of the device 100 is not confirmed.

[0091] In one particular embodiment, in order to compare the calculations performed with the expected results to confirm or refute the identity of the user of device 100, the comparison step may include a score calculation (step 224 in the diagram in Figure 3). The value of this score depends on the correlations between, on the one hand, the calculated position and geometric relationships, and on the other hand, the theoretical position and geometric relationships expected to confirm or refute the user's identity. For example, the score calculation method may yield a score close to or equal to 1 in a case such as the one previously described in relation to Figure 6, and a score close to 0 in cases such as those previously described in relation to Figures 7 and 8.

[0092] As an example, such a score can be calculated for each minutiae and point of interest determined previously. Then, it is possible to select some or all of the scores obtained and calculate an overall score representative of the previously calculated scores (sum, average, linear combination, etc.). The scoring method is not described in detail here and may correspond to an adaptation of one of the methods known to those skilled in the art, such as those used for fingerprint comparison.

[0093] The score value is then compared with a threshold value to assess the correspondence between the biometric measurement performed and the expected biometric data, and thus confirm or not the identity of the user (step 226).

[0094] In the embodiment described above, the features of the footprint of element 101 determined from the surface image 110 correspond to the minutiae 112 and 113 of this footprint, and the features of the microvasculature of element 101 determined from the overall volume image correspond to the microvasculature interest points 116. According to another embodiment, the determination of the features of the footprint and microvasculature of the element may involve extracting at least one feature vector from the surface image and the overall volume image using a deep learning algorithm.In this case, comparing the previously determined characteristics with expected theoretical characteristics to identify or reject element 101 may involve a comparison, using the deep learning algorithm, of the previously determined characteristic vector with at least one expected theoretical characteristic vector. In this alternative example, it is also possible to calculate a score whose value depends on the correlations between the determined characteristic vector and the expected one for identifying or rejecting element 101.

[0095] The document Engelsma, Joshua J. et al. "Learning a Fixed-Length Fingerprint Representation." IEEE Transactions on Pattern Analysis and Machine Intelligence 43 (2019): 1981-1997, describes examples of deep learning algorithms that can be implemented for this other example of Implementation. For example, in this document, Section 3 describes a CNN (convolutional neural network) deep learning algorithm for extracting features from a fingerprint, the architecture of the convolutional neural network used being shown in Figure 8 of this document. By training this algorithm on a fingerprint database as described in Section 7 (composed of Nv views of Ni identities, for example Ni fingers, different), an algorithm is obtained that projects views of the same identity close to each other in an Nd-dimensional space and far from other identities in the same space, so that a distance calculation in this space provides a relevant matching algorithm, as described in Section 4, Formula 11 of this document.It is possible to apply the method described in this document to data acquired by the biometric identification device 100, in this case, surface images superimposed and aligned with volume images. Furthermore, it is possible to adapt the convolutional neural network architecture to the fact that the input images are 3D. In addition, a training database consisting of surface and volume images with multiple identities, for example, multiple fingers (more than 100), and multiple views of each identity (at least 20), can be used. The resulting algorithm is a CNN-type algorithm that produces, from a single biometric input view, a feature vector of size Nd (with the coordinates of the projection in Nd-dimensional space).Next, to compare two vectors, the distance algorithm presented in this document can be applied and produce a matching score as described previously.

[0096] The paper by Y. Tang et al., "FingerNet: An Unified Deep Network for Fingerprint Minutiae Extraction", arXiv: 1709.02228, also describes examples of algorithms Deep learning can be implemented for this other embodiment. This document describes, in particular, a method for extracting fingerprint minutiae using deep learning. A similar method can be adapted to fingerprint images obtained with the 100 biometric identification device to extract minutiae. The learning method and an internal database can be used for the implementation of the method. An adaptation of this method can also enable the extraction of points of interest in volume images, for example, by manually annotating the points of interest in the images in a training dataset. This adaptation can consist of modifying the neural network architecture so that it takes 3D images as input. Another adaptation can enable the construction of a single algorithm that extracts minutiae from surface images and points of interest from volume images.

[0097] Device 100 can thus form a complete biometric system integrating identification, anti-spoofing, and liveness detection functionalities, based on a single compact representation of information from the captured surface imprint and also from microvascular information. Device 100 can notably enable:

[0098] - a better level of security compared to an identification device based solely on a fingerprint capture;

[0099] - a better level of assurance of anti-spoofing and detection of living beings due to the reliability of the data relating to microvasculature.

[0100] In the example described above, device 100 performs biometric identification by comparing the measured data included in the surface image and the image of volume with expected data corresponding to that of a person whose identity is verified. Alternatively, it is possible for device 100 to compare the measured data with expected data corresponding to that of several people, in which case device 100 can verify if the user corresponds to one of these people.

[0101] In the description above, device 100 can compare measurements taken with expected data stored in its memory. Alternatively, device 100 can connect to a remote database containing expected data corresponding to one or more individuals to verify if the device user matches that individual or individuals. Another alternative is to connect device 100 to a smart card containing data recorded during prior enrollment. The data comparison step can then be performed on the smart card rather than on device 100.

[0102] In the description above, device 100 is used to directly confirm or deny the identity of the user of device 100. Prior to this identification, device 100 can be used to perform an initial data acquisition (surface image(s) and volume images) during a prior enrollment, which will serve as expected data during future identifications of the user of device 100. Other uses of device 100 may also be envisaged.

[0103] Various embodiments and variations have been described. Those skilled in the art will understand that certain characteristics of these various embodiments and variants could be combined, and other variants will appear to the person of the trade.

[0104] Finally, the practical implementation of the described methods and variants is within the reach of the person in the trade, based on the functional indications given above.

Claims

CLAIMS 1. Biometric identification device (100) comprising at least one ultrasonic transducer circuit (104), a control circuit (106) and a data processing circuit (108), configured to implement at least the following steps: - acquisition (206, 210) of at least one surface image (110) of at least one element (101) of the body of a user of the device (100) intended to be placed against a capture surface (102) of the device (100), and of at least one series of volume images of the element (101) from which a global volume image (114) of a microvasculature of the element (101) is determined; - determination (214, 216, 218, 220) of characteristics of an imprint of the element (101) from the surface image (110), and of characteristics of the microvasculature of the element (101) from the global volume image (114); - comparison (222, 224, 226) of the characteristics determined in the previous step (214, 216, 218, 220) with characteristics calculated during a preliminary step of enrolling the user to identify or not the element (101); in which the data processing circuit (108) is configured to implement, between the step of acquiring (206, 210) the surface image (110) and the series of volume images and the step of determining (214, 216, 218, 220) the characteristics of the imprint and the microvasculature of the element (101), a step of aligning (212) the surface image (110) and the global volume image (114) with respect to each other.

2. Biometric identification device (100) according to claim 1, wherein the data processing circuit (108) is configured to determine the global volume image (114) by implementing a spatio-temporal filtering method of the series of volume images of the element (101).

3. Biometric identification device (100) according to one of the preceding claims, in which the ultrasonic transducer circuit (104) is configured to carry out the acquisition of the series of volume images of the element (101) by pulsed Doppler type imaging.

4. Biometric identification device (100) according to one of the preceding claims, in which the data processing circuit (108) is configured such that the step of determining (214, 216, 218, 220) the characteristics of the imprint and the microvasculature of the element (101) comprises at least: - determination (214, 216) of minutiae (112, 113) of the imprint of the element (101) from the surface image (110), and of points of interest (116) of the microvasculature of the element (101) from the global volume image (114); - calculation (220) of the position of at least one minutia (113) in the surface image (110) and / or of at least one point of interest (116) in the global volume image (114), and of geometric relationships between said at least one minutia (113) and / or said at least one point of interest (116) and at least a part of the other minutiae (112) and points of interest (116) taking into account, for the points of interest (116), the depths at which the points of interest (116) are located in the global volume image (114); and such that the comparison step (222, 224, 226) comprises a comparison of the previously calculated position and geometric relationships with a theoretical position and geometric relationships expected to identify or not the element (101).

5. Biometric identification device (100) according to claim 4, wherein the points of interest (116) correspond to bifurcations of micro-blood vessels and / or to points of strong curvature of micro-blood vessels.

6. Biometric identification device (100) according to one of claims 4 or 5, wherein the data processing circuit (108) is configured to implement, between the step of determining (214, 216) the minutiae (112, 113) and the points of interest (116) and the calculation step (220), a step of combining (218), in the form of a point cloud, the minutiae (112, 113) and the points of interest (116) determined.

7. Biometric identification device (100) according to one of claims 4 to 6, wherein the data processing circuit (108) is configured such that, during the calculation step (220), the other minutiae (112) and points of interest (116) with which the geometric relationships are calculated correspond to the minutiae (112) included in a region of the surface image (110) delimited by a planar geometric figure and to the points of interest (116) included in a region of the global volume image (114) included in a volume (120) formed from the planar geometric figure which is extended in a dimension of the global volume image (114) perpendicular to the surface image (110).

8. Biometric identification device (100) according to claim 7, wherein the planar geometric figure (118) corresponds to a polygon, a circle or an ellipse.

9. Biometric identification device (100) according to one of claims 7 or 8, the data processing circuit (108) is configured such that the calculation step (220) is repeated for each minutia (112, 113) and each point of interest (116) determined, and that the comparison step (222, 224, 226) is carried out for the position and the geometric relationships calculated for each minutia (112, 113) and each point of interest (116).

10. The biometric identification device (100) of claim 4, wherein the data processing circuit (108) is configured such that the calculated geometric relationships between a first minutiae (113) or point of interest (116) and a second minutiae (112) or point of interest (116) include the distance between the first minutiae (113) or point of interest (116) and the second minutiae (112) or point of interest (116), an orientation of the first minutiae (113) or point of interest (116) relative to the second minutiae (112) or point of interest (116), and an orientation of the second minutiae (112) or point of interest (116) relative to the first minutiae (113) or point of interest (116).

11. Biometric identification device (100) according to one of claims 4 to 10, wherein the data processing circuit (108) is configured such that, for a point of interest (116) corresponding to a bifurcation or a point of strong curvature of micro-blood vessels, the calculated geometric relationships include the orientations of each of the micro-blood vessels in relation to other minutiae and points of interest.

12. Biometric identification device (100) according to one of claims 1 to 3, wherein the data processing circuit (108) is configured such that: - the determination (214, 216, 218, 220) of the characteristics of the imprint and of the microvasculature of the element (101) comprises the extraction of at least one vector of characteristics of the surface image (110) and of the global volume image (114) by the implementation of a deep learning algorithm; - the comparison (222, 224, 226) of the previously determined characteristics with expected theoretical characteristics to identify or not the element (101) comprises a comparison, by the deep learning algorithm, of the previously determined characteristic vector with at least one expected theoretical characteristic vector.

13. Biometric identification device (100) according to one of the preceding claims, in which the data processing circuit (108) is configured such that the comparison step (222, 224, 226) comprises a calculation (224) of a score whose value depends on the correlations between the determined characteristics and the theoretical characteristics expected to identify or not the element (101).

14. Biometric identification device (100) according to one of claims 4 to 11 and according to claim 13, wherein the data processing circuit (108) is configured such that, during the comparison step (222, 224, 226), a score is calculated for each minutia (112, 113) and each point of interest (116) determined.

Images