Passive 3D object authentication based on image size

Abstract

A passive three-dimensional (3D) object authentication based on image size is described, for example for biometric facial recognition. For example, during an enrollment routine, an imaging system captures images of an enrollment user's face from multiple distances. The images can be processed to extract enrollment dimensions, including a single deterministic structural dimension, dimensional relationships that are static with changes in imaging distance, and dimensional relationships that predictably change with changes in imaging distance. During an authentication routine, the imaging system again captures an authentication image of an authenticating user's face (purportedly a previously enrolled user) at an authentication imaging distance and processes the image to extract authentication dimensions. Expected and actual dimensional quantities are calculated from the authentication and enrollment dimensions and compared to determine whether the authenticating user's face is authorized as previously enrolled and / or whether it is a spoof.

Description

[0001] This application claims priority to U.S. Application No. 17 / 352,375, filed June 21, 2021, entitled “Passive Three-Dimensional Object Authentication Based on Image Size Ratios,” the entire contents of which are incorporated herein by this reference. TECHNICAL FIELD

[0002] Embodiments generally relate to optical devices integrated into personal electronic devices. More specifically, embodiments relate to passive three-dimensional object authentication based on image size ratios, such as for biometric facial recognition and spoof detection. BACKGROUND

[0003] Many modern electronic devices, such as smartphones, tablets, and laptops, are equipped with biometric security access systems, such as facial recognition (ID), fingerprint sensors, and the like. For example, facial recognition can be used to unlock a smartphone, log into applications and accounts, authorize mobile payments, and the like. Similar facial recognition technology is integrated into other access control devices, such as electronic locks and automated teller machines (ATMs). Effective implementation designs often balance various considerations. For example, it is often desirable to provide fast and accurate results to users in a manner that avoids false positives (which can degrade the security of the implementation) and false negatives (which can frustrate authorized users).

[0004] Traditional facial recognition systems often include relatively fast, but relatively insecure, facial recognition methods that are based on recognition of a limited number of large-scale structures. These methods are often relatively easy to spoof, for example, by using a two-dimensional image of an authorized user’s face, a three-dimensional wax or latex model of an authorized user’s face, and the like. For example, traditional facial recognition implementations on smartphones are often designed to minimize the use of battery resources, memory resources, processor resources, and the like. In addition, traditional facial recognition implementations on smartphones often do not place a high emphasis on advanced spoofing techniques and the like; moreover, to avoid frustrating authorized users who are trying to quickly unlock a smartphone, they often allow more false positives than false negatives. However, for many smartphones and other applications, it can be desirable to provide a higher level of security (including additional anti-spoofing protection) without unduly impacting battery, memory, processor, and other resources. SUMMARY

[0005] Embodiments provide image size-based passive three-dimensional (3D) object authentication, such as for biometric facial recognition. For example, during an enrollment routine, an imaging system (e.g., in a smartphone or other electronic device) captures images of an enrolled user’s face from multiple distances. The images can be processed to extract enrollment dimensions, including individual deterministic structure dimensions, dimension relationships that are static as the imaging distance changes, and dimension relationships that change predictably as the imaging distance changes. During an authentication routine, the imaging system again captures an authentication image of an authenticating user’s face (purportedly of a previously enrolled user) at some authentication imaging distance, and processes the image to extract authentication dimensions. Expected and actual dimension quantities are computed from the authentication and enrollment dimensions and compared to determine whether the authenticating user’s face is authorized as previously enrolled and / or is a spoof.

[0006] According to one set of embodiments, a method for authenticating an authenticating object based on passive imaging is provided. The method includes capturing a set of images of the authenticating object using an imaging system, the imaging system being at an authentication imaging distance from a first deterministic structure of a plurality of deterministic structures of the authenticating object visible in the set of images; processing the set of images to measure at least a first authentication dimension of the first deterministic structure and a second authentication dimension of a second deterministic structure of the plurality of deterministic structures; obtaining enrollment dimensions from an enrollment database, the enrollment database being obtained by measuring at least the first deterministic structure and the second deterministic structure in a previous enrollment process; computing the authentication imaging distance based on a relationship between the first authentication dimension and the enrollment dimensions, and computing an expected imaging distance for the second deterministic structure based on a relationship between the authentication imaging distance and the enrollment dimensions; computing an authentication normalization factor as a ratio between the first authentication dimension and the second authentication dimension, and computing an expected normalization factor for the second deterministic structure at the expected imaging distance based on the enrollment dimensions and the authentication imaging distance; and authenticating the authenticating object based on comparing the authentication normalization factor to the expected normalization factor. In some such embodiments, the method further includes outputting a user-perceptible authentication result to indicate whether the authentication result is an authorization or a denial of authentication.

[0007] According to another set of embodiments, a system for authenticating an authentication subject based on passive facial imaging is provided. The system includes: an enrollment database storing a plurality of enrollment dimensions obtained by measuring a plurality of deterministic structures of the authentication subject during a previous enrollment process; an imaging subsystem for capturing a set of images of the authentication subject, wherein the authentication subject is at an authentication imaging distance from a first deterministic structure of the plurality of deterministic structures; and a control and processing module having one or more processors and a memory storing a set of instructions that, when executed, cause the one or more processors to perform steps. The steps include: processing the set of images to measure at least a first authentication dimension of the first deterministic structure and a second authentication dimension of a second deterministic structure among the plurality of deterministic structures; calculating the authentication imaging distance based on a relationship between the first authentication dimension and the enrollment dimension; calculating an expected imaging distance for the second deterministic structure based on a relationship between the authentication imaging distance and the enrollment dimension; calculating an authentication normalization factor as a ratio between the first authentication dimension and the second authentication dimension; calculating an expected normalization factor for the second deterministic structure at the expected imaging distance based on the enrollment dimension and the authentication imaging distance; and authenticating the authentication subject based on comparing the authentication normalization factor with the expected normalization factor. In some such embodiments, the system further comprises an output interface in communication with the control and processing module to output a response and instruct the control and processing module to authenticate the authentication object and output a user-perceivable authentication result. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0009] Figure 1 A mobile device with an integrated imaging system is shown.

[0010] Figure 2 A simplified block diagram of an electronic access control system using the registration and authentication techniques described herein to authorize or deny access is shown in accordance with some embodiments.

[0011] Figure 3 An imaging environment is shown for illustrating the optical principles involved in deriving depth information from imaging using image size as a background for the embodiments herein.

[0012] Figure 4 Illustrative images of faces with various deterministic macrostructures are shown.

[0013] Figure 5An illustrative set of imaging data is shown as background to various embodiments described herein.

[0014] Figure 6 An illustrative partial chromatic response is shown for an illustrative portion of a feature sub-image.

[0015] Figure 7 An illustrative facial image is shown with one or more calibrators placed in the imaging region.

[0016] Figure 8 A flowchart of an illustrative method for enrolling an authentication object, such as a user’s face, according to various embodiments herein is shown.

[0017] Figures 9A-9D An ever-changing image magnification in example images acquired by imaging an experimental setup at different imaging distances is shown.

[0018] Figure 10 A flowchart of an illustrative method for authenticating an authentication object by passive imaging of a three-dimensional deterministic structure of the authentication object according to various embodiments herein is shown.

[0019] Figure 11 A flowchart of an example gating process for access control according to various embodiments is shown.

[0020] Figure 12 A schematic diagram of one embodiment of a computer system that can implement various system components and / or perform various steps of methods provided by various embodiments is provided.

[0021] In the appended figures, similar components and / or features can have the same reference label. Further, various components of the same type can be distinguished by following the reference label by a dash along with a second label that distinguishes among the components of the same type. If only the first reference label is used in the specification, the description is applicable to any one of the similar components having the same first reference label irrespective of the second reference label. DETAILED DESCRIPTION

[0022] In the following description, for the purposes of explanation, numerous specific details are set forth in order to thoroughly understand the application. It will be apparent, however, to one skilled in the art that the application can be practiced without one or more of these details. In other instances, well-known features and techniques have not been described in order to simply the present description.

[0023] Many modern electronic devices have integrated imaging systems that can be used for a variety of functions. For example, integrated imaging systems are ubiquitous in smartphones, automated teller machines, physical access control systems (e.g., electronic door locks), and the like. In some cases, such imaging systems can provide user identity verification functions, e.g., for access control, biometric verification, and the like. Some imaging-based identity verification functions make use of face ID. For example, face ID can be used to provide depth and / or focus information to the same and / or other imaging systems to verify the authenticity or identity of a user, and / or for other purposes.

[0024] For purposes of context, Figure 1 A mobile device 100 having an integrated imaging system is shown. According to some embodiments, the mobile device can be a smartphone, tablet, laptop, and the like. The mobile device 100 can include a display screen 120, a frame 110 surrounding the display screen 120, control buttons (e.g., a power button 140, a volume control button 130, a grip force sensor 150, and the like), and / or any other components. The integrated imaging system can include one or more cameras 160, e.g., a front-facing camera (e.g., a "selfie" camera), a rear-facing camera, and the like. In some implementations, the front-facing camera 160 is also used for face ID sensing. In other implementations, a dedicated camera 160 is provided for face ID sensing.

[0025] Figure 2 A simplified block diagram of an electronic access control system 200 that uses the enrollment and authentication techniques described herein to authorize or deny access according to some embodiments is shown. An imaging camera 210, e.g., a camera 160 in Figure 1 The camera 210 can have an associated field of view (FOV) within which it can image a three-dimensional (3D) object 202. In the context described herein, it is generally assumed that the 3D object 202 is a face, or purported to be a face. However, it should be understood that such description should not be construed as limiting the present application to operating only in the context of faces and / or face ID sensing; rather, the embodiments described herein can be applied in any suitable imaging context.

[0026] The camera 210 uses imaging optics (e.g., lenses, mirrors, filters, and the like), a sensor (e.g., a photodetector), and / or any suitable components to capture an image 220 of the 3D object 202. In some embodiments, capturing the image 220 can involve focusing the imaging optics and / or adjusting the sensor to form a sharp image having desired acuity, contrast, chromaticity characteristics, and the like. For example, the captured image 220 can be free of or have very little distortion or other types of image aberrations (e.g., spherical aberration, coma, astigmatism, and field curvature).

[0027] The control and processing module 230 can analyze features of the image 220. For example, the control and processing module 230 can be used to identify that the image 220 contains an image of a human face, and can extract facial features of the human face from the image 220. The control and processing module 230 can also be used to compare the facial features in the image 220 to facial features of authorized users stored in a registration database 240 (e.g., one or more registration databases). The registration database 240 can include facial recognition data of authorized users generated during a registration process. For example, during the registration process, one or more images of a live face of an authorized user can be captured by the camera 210. The image can be analyzed to extract (e.g., and characterize, etc.) facial features of the authorized user. The facial image of the authorized user, as well as the facial features, can be stored in the registration database 240 for subsequent security checks.

[0028] The control and processing module 230 can determine whether the facial features in the ideal image 220 match the facial recognition data stored in the registration database 240. The control and processing module 230 can output a facial recognition decision through the output interface 250. The electronic access control system 200 can grant or deny access based on the decision provided through the output interface 250. For example, the electronic access control system 200 can be used to grant physical access (e.g., as part of a biometric door lock, a biometric lock for a safe or file cabinet, etc.), logical access (e.g., providing access to a mobile device, a computing system, an automated teller machine account access, etc.), and / or any other type of access control. For example, the processing unit of the mobile device 100 can grant or deny access based on the facial recognition decision, and / or provide other features. For example, if the control and processing module 230 outputs a positive facial recognition decision indicating a match, the processing unit of the mobile device 100 can grant access to the mobile device 100 (e.g., wake up the mobile device 100), grant use of the mobile device for a payment transaction, etc. Figure 1 On the other hand, if the control and processing module 230 outputs a negative facial recognition decision indicating a mismatch, the processing unit of the mobile device 100 can deny access to the mobile device 100 (e.g., the mobile device 100 can remain locked).

[0029] In many practical contexts, such facial recognition sensing aims to balance potentially competing considerations. For example, conventional approaches tend to analyze captured images 220 only to the extent that a relatively small number of feature points or measurements can be extracted from large-scale facial structures (e.g., eye corners). Those extracted points or measurements are then compared to previously registered points or measurements to determine whether a statistical match is apparent. Such an approach can be relatively fast and lightweight (e.g., using minimal computation, memory, battery, and / or other resources), but can also provide a relatively low level of security. For example, such an approach can produce false positive matches for individuals with similar appearances, and / or can be relatively easy to spoof using high-resolution two-dimensional images of authorized user faces, three-dimensional wax or latex models of authorized user faces, and so on. For many smartphones and other applications, it can be desirable to provide a higher level of security (including additional anti-spoofing protection) without unduly impacting battery, memory, processor, and other resources.

[0030] Generally speaking, embodiments described herein exploit various optical principles, including image size, to derive depth information from imaging. Figure 3 An imaging environment is shown that illustrates optical principles involving the use of image size to derive depth information from imaging, as background to embodiments herein. As shown, a lens 310 forms an image 325 of an object 320 on a photodetector 305 (e.g., a photodiode array). The lens 310 has a known focal length (f) 315, the photodetector 305 is a known imaging distance (r) 335 from the lens 310, and the object 320 has a known object dimension (H) 340. When the object 320 is a particular object distance (R) 330 from the lens 310, the image 325 will exhibit a dimension (h) 345 that corresponds to H 340.

[0031] In such a configuration, R 330 is a function of f 315 and H 340 in the context of a known value of h 345 (and, in some cases, r 335 as well). For example, R 330 can be determined by measuring h 345 (e.g., measuring several pixels in dimension h 345) according to the following calculation: R ~ f * H / h. As described herein, in cases where object 320 includes multiple structures, multiple such measurements and calculations can be used to obtain multiple distance measurements. In some embodiments, a particular dimension of a particular structure can be used as a unit dimension length, etc.; and other features can be determined in relation to that unit dimension length (referred to herein as "feature relative dimensions"). For example, the width of a user's left eye is used as a unit dimension length (i.e., the user's left eye is 1.000 left eye widths), and other facial dimensions of the user are determined in relation to the same unit (e.g., the same user's mouth is measured to be 1.328 left eye widths). In other embodiments, a ruler and / or other calibration method is used to establish dimensions in non-relative dimension units such as millimeters, number of pixels, etc. (referred to herein as "absolute dimensions").

[0032] In the context of human face recognition sensing, many deterministic macrostructures (i.e., large-scale structures) have been shown to provide measurements of relative features of a particular individual, and tend to remain consistent for that particular individual over time. Figure 4 An illustrative image 400 of a face with various deterministic macrostructures is shown. As shown, the deterministic macrostructures generally correspond to the perceived organs of the face, such as the eyes, nose, and ears. For example, the illustrated image 400 indicates measurements of eye width 410 (e.g., between the outer corners of the eyes), distance between eyes 420 (e.g., between the inner corners of the two eyes), iris diameter 415, distance from the corner of the eye to the bottom of the nose 425, height of the ear 430, width of the nose 435, etc. Additional deterministic macrostructures can be used in some implementations, such as size of the front teeth, length of the nose, etc.

[0033] When certain macrostructures are not sufficiently determinative, they can be excluded from use in a facial recognition sensing context. For example, the mouth can change shape significantly with changes in mood and facial expression, and thus the mouth can not provide sufficiently determinative macrostructure measurements. Similarly, the eyebrows, forehead wrinkles, hairline, pupil diameter, and other large-scale structures in a facial image can vary from one imaging session to another. In addition, some determinative macrostructures tend to be selected or excluded in facial recognition sensing contexts based on how easily or reliably they can be measured. For example, a conventional facial recognition sensing technology can not reliably locate the tip of the nose because that location can not have easily identifiable features, or a conventional facial recognition sensing technology can not reliably locate the tip of the earlobe because the earlobe can not reliably be in the field of view of the imaging system.

[0034] Different implementations and implementation contexts can yield different approaches to obtaining determinative macrostructure measurements. When a user turns or tilts her head relative to the imaging system, and / or changes her distance from the imaging system, certain measurements can change. Nonetheless, the impact on the measurements tends to be largely determinative. For example, when the head is turned, each eye is at a different distance from the imaging system and has a different 3D orientation relative to the imaging system; however, these eye-to-eye differences tend to follow a predictable mathematical pattern. Thus, rather than simply relying on determinative macrostructure measurements obtained directly from the image, embodiments can employ additional calculations, standardizations, statistical processes, and / or other processes to account for these and other types of impact on the measurements. For example, in some implementations, the determinative macrostructure measurements include calculated measurements. For example, one implementation can measure the distance between the eyes 420 and the distance from the corners of the eyes to the bottom of the nose 425, and can further calculate the ratio of the measurements 420 and 425 as a calculated measurement. In some embodiments, this calculated measurement is used as part of a set of determinative macrostructure measurements. In other embodiments, this calculated measurement is used to correct a set of determinative macrostructure measurements. For example, an expected determinative macrostructure measurement can be considered to be in an expected measurement plane, and any change in the orientation or position of the user's head effectively changes the orientation and position of the measurement plane to the orientation and position of the imaging measurement plane (and correspondingly changes the position and orientation of the obtained determinative macrostructure measurements). The calculated measurement can be used to mathematically characterize the orientation and / or position of the imaging measurement plane, and to determine and apply a corresponding mathematical transformation to reposition and orient the obtained determinative macrostructure measurements to the expected measurement plane.

[0035] Figure 5An illustrative set of imaging data 500 is shown that is illustrative of the background of various embodiments described herein. The illustrative set of imaging data 500 includes a high-definition image 510 of a portion of a human face, as well as a plurality of feature sub-images 520, each associated with a respective deterministic microstructure feature region. The illustrative set of imaging data 500 is merely intended to illustrate features of embodiments and is not intended to limit the types of images described herein. For example, while the illustrative set of imaging data 500 includes processed output images from an imaging system, some embodiments described herein rely on imaging data that includes raw output data from an image sensor (e.g., uncorrected for color or otherwise processed).

[0036] The "subject" (i.e., the imaged portion of the human face) shown in image 510 includes many different types of traceable structures. As described herein, embodiments can use these traceable structures to reliably locate deterministic microstructure feature regions. In some embodiments, for example Figure 4 As described, traceable structures are deterministic macrostructures. For example, traceable structures include eyes, ears, and a nose. Such deterministic macrostructures can yield deterministic measurable feature locations, such as a corner of an eye, an eye width, a nose width, etc. In some embodiments, traceable structures can include additional macrostructures that are not necessarily deterministic, but can still be reliable locators of deterministic microstructure feature regions. For example, traceable structures in such embodiments can include eyebrows, eyelashes, eyelids, nostrils, lips, etc.

[0037] Deterministic microstructures can be small-scale structures of an imaged subject that are consistent enough from imaging session to imaging session to be used for face recognition. Without additional image processing, such deterministic microstructures tend to not be easily seen or characterized. In some embodiments, such deterministic microstructures are skin texture features, such as a pore pattern. Notably, deterministic microstructures do not include non-deterministic features. For example, a freckle pattern can tend to change over time, such as with a person's recent sun exposure, etc. In some implementations, deterministic microstructures can include other types of small-scale deterministic structures, such as an iris vein pattern, etc. However, some such microstructures, even if deterministic, can still be susceptible to spoofing. For example, a high-resolution photograph can sufficiently capture a vein pattern in a person's eye to spoof an imaging system (e.g., as opposed to skin texture, which even a highest resolution photograph can not be able to capture). Accordingly, some embodiments avoid using those types of deterministic microstructures for face recognition sensing, or use only those types of deterministic microstructures along with other types of deterministic microstructures that are less susceptible to spoofing.

[0038] Deterministic macrostructure measurements can be used to locate deterministic microstructure feature regions in any suitable manner. For example, as illustrated, deterministic macrostructure measurements can be used to generate various grids, distances, angles, etc. from which to guide the location of one or more deterministic microstructure feature regions. As one example, a first deterministic microstructure feature region is known to be located at a certain vertical distance from the bottom of the nose (e.g., based on prior registration). When imaging the face, a vertical reference is computed as the center between the eyes to the center of the chin; and the first deterministic microstructure feature region can be found at a corresponding distance from the bottom of the nose along this vertical reference. Thus, a first feature sub-image 520a can be derived to correspond to the first deterministic microstructure feature region. As another example, a second deterministic microstructure feature region is known to be located on the cheek at a particular intersection of a reference line and a triangle. Specifically, when imaging the face, the vertices of the triangle are located at the bottom center point of the nose, the outer corner of the right eye, and the center point of the chin; a horizontal reference line is located at a position that passes through the bottom center point of the nose and is perpendicular to the vertical reference line; and the location of the second deterministic microstructure feature region is derived from the intersection of the horizontal reference line and the hypotenuse of the triangle. Thus, a second feature sub-image 520b can be derived to correspond to the second deterministic microstructure feature region.

[0039] After the deterministic microstructure feature regions are located, corresponding feature sub-images 520 at these locations can be processed to derive deterministic microstructure measurements. Figure 6 Illustrative portions of illustrative portions chromatic responses 600 are shown for two feature sub-images. In particular, a first chromatic response 600a is derived from a portion of the feature sub-image 520d in Figure 5 corresponding to a feature region around the tip of the nose; and a second chromatic response 600b is derived from a portion of the feature sub-image 520b in Figure 5 corresponding to a feature region around the cheek. Each illustrated chromatic response 600 is a plot of luminance values of a chromatic component (e.g., signal levels of a corresponding photodetector element) on a 50-pixel long row of the image sensor array versus position. For example, the imaging system is focused using a mid-wavelength chromatic component, e.g., using a "green" chromatic response, and the plotted values indicate the response amplitude of the red photodiode component of each pixel of the photodiode array of the imaging system. Some embodiments described can operate regardless of which color component is used for focusing, color correction, imaging, chromatic response plots, etc. For example, certain chromatic configuration options can tend to increase response amplitude, improve contrast, and / or otherwise improve certain sensing parameters; however, some embodiments can still reliably operate across a large range of such chromatic configuration options.

[0040] A variety of types of information can be obtained from the chrominance response 600. To obtain such information, implementations can compute statistics to measure the distribution of luminance slopes, the standard deviation of luminance valley depths, the profile valley widths, and / or other values. For example, an illustrative luminance valley depth 620 and an illustrative profile valley width 610 are shown in each plot 600. Valley depth information can be expressed in terms of image contrast, average valley width at average valley depth can be computed, forming a face structure plot. Such plots and values can be generated and analyzed (i.e., computed) across portions or all of the image. For example, some implementations compute these values across the entire facial image, while other implementations compute these values only within predefined and located areas of certain microstructure features. As described above, various plots and values can be mapped to facial locations according to certain macrostructure measurements and locations. In some embodiments, the derived certain microstructure measurements are mapped to certain macrostructure locations to establish a 3D plot of the entire face or face portion.

[0041] It can be seen that the chrominance response plots 600 can be used to obtain (e.g., extract, derive, compute, etc.) texture features. The pattern of peaks and valleys in the chrominance response 600 can be indicative of certain microstructure, e.g., pores and / or other texture variations of the skin in the respective portions of the respective feature sub-image 520. Obtaining and characterizing such texture features can support certain features. One such feature is that the sensing of the presence of such texture features clearly indicates that the imaged object is a 3D object with a microstructure pattern. For example, a high-resolution 2D photograph can match certain macrostructure of a pre-registered individual. However, such a photograph will not include such microstructure, and the imaging of the photograph will not produce such texture features. Notably, such functionality does not rely on the pre-registration or matching of any particular texture feature; only the presence of texture features. For example, one implementation can use certain macrostructure for face recognition sensing, and can further detect the presence of any texture features to ensure that the imaged object is not a 2D photograph, or a 3D model without texture.

[0042] Another use of such texture features is to determine whether the obtained texture features are characteristic of the imaged object. For example, the pattern and / or range of measured valley depths 620 and valley widths 610 obtained from a facial skin image can tend to be relatively consistent across most or all human faces. Thus, some embodiments can determine whether the derived texture features are characteristic of a human face, even without pre-registering or matching any particular texture feature. For example, some such embodiments can use certain macrostructure for face recognition sensing, and can further detect the presence of characteristic texture features to indicate that the imaged object is a real human face (i.e., not a 2D photograph, 3D model, or other deception).

[0043] Another use of such texture features is to determine whether the acquired texture features are of a particular, pre-registered user. For example, a particular pattern or measurement of valley depth 620 and valley width 610 acquired from a feature sub-image 520 of a user's face corresponds to a deterministic microstructure that is uniquely the user's (e.g., unique enough for use in facial recognition, user authentication, biometric verification, etc.). Thus, some embodiments can determine whether the derived texture features match a profile set of (i.e., pre-registered) texture features of the user purported to be imaging. In such embodiments, the facial recognition sensing can use the deterministic macrostructure and the deterministic microstructure to support verification of user identity and spoof detection.

[0044] While the illustrated graphs show only a single chrominance component (i.e., red) chrominance response graph, chrominance information can yield additional information that can support additional features. In some embodiments, a single narrowband light wavelength is used for facial recognition sensing. For example, a particular wavelength is selected to yield a sharp contrast in chrominance response across a wide range of skin tones, pigments, and other features. Some embodiments can use light within the visible spectrum. Other embodiments can additionally or alternatively use light outside the visible spectrum, such as light in the infrared (e.g., near infrared) or other spectrum. In some embodiments, relative and / or absolute depth information can be acquired by comparing chrominance response data across multiple chrominance responses. For example, an imaging system can be focused according to a green chrominance component, and can generate chrominance response graphs for red and blue chrominance components derived from imaging of an object at a distance from the imaging system. Because lenses tend to have different focal lengths for different chrominance components at different object distances, the difference in sharpness indicated by different color components in response to a particular feature sub-image 520 can indicate the relative (e.g., if in a calibration situation, or absolute) distance of the deterministic microstructure feature region corresponding to that feature sub-image 520. This depth information can be used for various purposes, such as to help determine the orientation and / or position of a face as an image, to help find one or more absolute reference distances, etc.

[0045] Some embodiments can be implemented with only relative distance measurements. As noted above, some such embodiments can rely on computationally-derived measurements or the like. For example, in imaging a face, the 2D or 3D coordinate locations of the bottom center point of the nose (A), the outer corner of the right eye (B), and the center point of the chin (C) can be recorded. In this case, all of these coordinate locations can be referenced to some generated image reference coordinate system, and can not have any relationship to absolute measurements. Nonetheless, a reference perpendicular can be generated from the straight line AC, a reference horizontal can be generated that intersects the reference perpendicular at point A, a reference triangle can be generated from triangle ABC, and the like. Without absolute distance measurements, the location of the deterministic microstructure feature region can be acquired from reference features. As one example, the pre-registered location of the deterministic microstructure feature region can be defined as the end of a vector that begins at point A and has a direction that bisects the straight line BC at some location (D) and extends 1.4 times the distance of AD.

[0046] In other embodiments, it can be desirable to acquire one or more absolute measurements. In some such embodiments, a calibrated chromatic differentiation can be used to derive at least one absolute depth measurement. In other such embodiments, measurements can be acquired in the context of reference measurement guides (e.g., at least during registration). To illustrate, Figure 7 An illustrative face image is shown with one or more calibrators 710 placed in the imaging region. The calibrators 710 can comprise any suitable reference measurement guide, such as a ruler, grid, frame, bar code, or anything of absolute size known to, or acquirable by, the imaging system. In some implementations, the calibrators 710 are implemented on a transparent substrate, such as a ruler printed on a transparent sticker. The calibrators 710 can be placed in one or more locations. For example, a subset of the deterministic macrostructure can be defined for registration and calibration, and the calibrators 710 (or multiple calibrators 710) can be placed in or near these locations during the registration process.

[0047] Figure 8A flowchart showing an illustrative method 800 for enrolling an authentication object, such as a user's face, in accordance with various embodiments herein is shown. At stage 804, an enrollment routine is activated. For example, facial recognition enrollment can be activated by an authorized user of a mobile device, such as by selecting facial recognition enrollment in "Settings," or when the mobile device is first opened by an authorized user upon purchase of the mobile device. While described with reference to facial recognition enrollment and the like, the process 800 can be used to enroll any suitable three-dimensional object having a deterministic measurable structure. Thus, reference to "facial recognition" should be understood to refer generally to any suitable three-dimensional structure authentication. As described herein, the enrollment process can involve capturing and processing images of the object at different imaging distances to establish a database of measurements of the macrostructure and / or microstructure of the target. Relationships between the measurements can be used for biometric authentication, spoof detection, biometric identification, and / or other purposes.

[0048] At stage 808, the user is prompted to establish a "far distance" imaging setup. In the described sequence, one or more "far distance" imaging setups are followed by one or more "near distance" imaging setups (e.g., starting at stage 824). However, other implementations can start with a near distance imaging setup, which can be followed by a far distance imaging setup. Still other imaging setups can involve multiple setups at multiple distances in any suitable order. Each of the one or more far distance setups can involve positioning an imaging device (e.g., a smartphone camera, a digital camera, a biometric access system camera, etc.) at a far distance that is relatively far from the object (e.g., the user's face). In some embodiments, the far distance is expected to be farther than a typical distance a user can seek to engage with a facial recognition system. For example, a user can typically use a smartphone facial recognition system while holding the smartphone about 20-40 centimeters from his or her face, and at stage 808 the user is prompted to hold the smartphone as far from his or her face as is comfortable (e.g., a distance of one arm length, such as about 45 centimeters). At stage 808, the user can also be prompted to orient his or her face in a particular way. For example, the user can be instructed to look directly at the camera, to focus on a particular point on the screen, etc.

[0049] At stage 812, with the user and imaging system in the far imaging setup, the camera can capture a set (i.e., one or more) of far distance images (e.g., far distance images of the user's face). For example, the set of far distance images can include one or more images under one or more imaging conditions, such as using one or more lighting conditions, one or more focus and / or zoom conditions, one or more cameras, one or more aperture or other optical settings, one or more orientations, etc. In some embodiments, at stage 810, the user can be instructed to place one or more calibrators (e.g., as described above with reference to FIG. 7) prior to capturing some or all of the images at stage 812. Figure 7The one or more images captured in stage 708 can include one or more of a reference measurement guide, such as a ruler, placed in the field of view of the camera. For example, a small ruler with known spacing (e.g., in millimeters) is placed on the user's nose, forehead, etc. in the context of a reference measurement guide, such as a ruler.

[0050] At stage 816, embodiments can process the image(s) to extract deterministic structural remote dimensions. Some of the extracted deterministic structural remote dimensions are clear observable measurements of large scale structural features of deterministic macrostructure. For example, an image of a user's face can be processed to extract deterministic macrostructure remote dimensions such as eye corner locations, eye width, nose width, nose height, interocular distance, etc. As noted above, some of the extracted deterministic structural remote dimensions can include measurements of small scale structural features of deterministic microstructure that can or can not be readily observable without additional image processing. For example, certain macrostructure remote dimensions and / or image processing techniques are used to identify regions of an image of an object (e.g., a face) that are most likely to include regions of deterministic microstructure features, such as relatively large skin regions that typically lack deterministic macrostructure or other traceable structures (e.g., skin regions of the cheeks, forehead, nose, etc.). Such embodiments can then determine a location definition (e.g., based on a set of reference points, reference lines, reference polygons, etc.) for each identified likely region. In some implementations, the deterministic structural remote dimensions (including macrostructure and / or microstructure dimensions) are defined based on a set of coordinates in a reference coordinate plane, which can be mathematically transformed (e.g., positioned, oriented, scaled, tilted, etc.) based on a subset of the measured dimensions. For example, a small number of highly deterministic macrostructure dimensions can be used to generate a transformation matrix, and other extracted dimensions are defined relative to the same coordinate system as those highly deterministic macrostructure dimensions, or in relation to those highly deterministic macrostructure dimensions. As noted above, some of the extracted deterministic structural remote dimensions can include measurements of small scale structural features of deterministic microstructure derived by computing deterministic microstructure from a feature sub-image. Statistical processing, and / or other image processing can be used to identify deterministic microstructure texture features in the image data. For example, average peak height, valley width, and / or other data can be extracted from the chrominance response data to indicate microtexture of a region, such as due to a pattern of pores, etc.

[0051] In some embodiments, at stage 818, one or more additional long-range imaging settings can be added to increase the amount of certainty structure long-range data. Embodiments can automatically change one or more imaging settings (e.g., focus, zoom, illumination, etc.), can prompt a user to manually change one or more imaging settings, and / or can prompt a user to change orientation relative to the imaging system. For example, embodiments can prompt a user to turn and / or tilt the head in one or more directions by one or more amounts. In each of the one or more additional long-range imaging settings, the method 800 can iterate back at least to capture additional long-range images at stage 812 and process the additional long-range images at stage 816. Different additional long-range imaging settings can tend to highlight certain structural features more or less (e.g., ears can be hidden in one orientation and visible in another, or certain microscopic texture features can only be clearly derived under certain illumination and / or zoom conditions), can affect the relationship between certainty structure longs, (e.g., the corners of the eyes can appear closer when the eyes are looking straight at the camera than when the eyes are slightly averted from the camera relative to the face), and / or can confirm certain certainty structure longs and / or relationships between certainty structure longs.

[0052] At stage 820, embodiments can establish a long-range enrollment database by storing the certainty structure longs and / or additional data computed therefrom (e.g., transformation matrices, etc.) in a data store. In embodiments supporting multiple users, the database can be established for a particular authorized user. For example, a face recognition database can store one or more enrollment profiles for one or more authorized users, and the stored certainty structure longs are associated with the enrollment profile of the particular authorized user that activated the enrollment process at stage 804.

[0053] At stage 824, the user is prompted to establish a "close-up" imaging setup. In some embodiments, stages 824, 828, 832, and 836 (e.g., in some embodiments, also 826 and / or 834) are implemented in substantially the same manner as stages 808, 812, 816, and 820 (e.g., in some embodiments, also 810 and / or 818), respectively, except that stages 824-836 are performed under one or more close-up imaging setups. Each of the one or more close-up setups can involve positioning an imaging device (e.g., a smartphone camera, a digital camera, a biometric access system camera, etc.) at a close-up distance that is relatively close to an object (e.g., a user's face). In some embodiments, the close-up distance is expected to be closer than a typical distance a user can seek to engage with a facial recognition system. For example, as noted above, a user can typically use a smartphone facial recognition system while holding the smartphone about 20-40 centimeters away from his or her face, and in stage 824 the user is prompted to hold the smartphone as far away from his or her face as possible while still allowing the imaging device to fully image the face, while still allowing the user to comfortably see the display screen (e.g., at about 15-20 centimeters). In some embodiments, the close-up distance is expected to be significantly different (e.g., at least 15 centimeters apart) from the far distance, etc. The user can be guided to select a suitable distance in any suitable manner. In one implementation, the user is generally instructed to hold the imaging device "at arm's length" (etc.) relative to stage 808, and the user is generally instructed to move the imaging device "as close to the face as possible while still fitting the entire face in the viewfinder" (etc.). In another implementation, a range finder (e.g., a time-of-flight sensor, a laser range finder, and / or other range finder) is used to provide feedback (e.g., graphical, audible, visual, and / or other suitable feedback) to the user to indicate whether the imaging device is in the appropriate relative position in stages 808 and 824. The same or different guidance can be used in stages 808 and 824. In stage 824, the user can also be prompted to orient his or her face in a particular manner, such as being instructed to look directly at the camera, to focus on a particular point on the screen, etc.

[0054] At stage 828, the camera can capture a set of close-up images with the user and imaging system in the close-up imaging setup. The images can be captured in stage 828 in the same manner as in stage 812.

[0055] In some embodiments, at stage 826, the user can be instructed to place one or more calibrators to provide an absolute measurement reference prior to capturing some or all of the images at stage 828. At stage 832, embodiments can process the images to extract deterministic structure proximities. The images can be processed at stage 832 in the same manner as in stage 816. For example, the extracted deterministic structure proximities can include measurements of large-scale structural features of deterministic macrostructure, measurements of small-scale structural features of deterministic microstructure, deterministic microstructure texture features in the image data, and / or any other suitable dimension. In some embodiments, at stage 834, one or more additional proximity imaging settings can be added to increase the amount of deterministic structure proximity data. For example, as described with reference to stage 818, embodiments can automatically change one or more imaging settings, prompt the user to manually change more than one imaging setting, and / or prompt the user to change orientation relative to the imaging system.

[0056] At stage 836, embodiments can establish a proximity registration database by storing the deterministic structure proximities and / or additional data computed therefrom (e.g., transformation matrices, etc.) in a data store. In some embodiments, the same data store (e.g., the same database) is used to store the proximity registration database and the distance registration database. For example, all of the deterministic structure distalities and the deterministic structure proximities (e.g., as well as any other data derived therefrom) are stored in a single registration database. As described above, the proximity registration database can be established for a particular authorized user that activates the registration process at stage 804, or the registration database can be combined.

[0057] After the registration database is established, various relationships can be established, and these relationships can be used to provide various features. Table 1 below provides some illustrative relationships that are obtainable using the deterministic structure distalities and the deterministic structure proximities:

[0058]

[0059]

[0060] The left half of Table 1 shows the relationship of the deterministic structure far dimension and two example structures: a reference structure (r), and a“junk” structure (j). The right half of Table 1 shows the relationship of the deterministic structure near dimension and the same two example structures. For example, imaging a user’s face, the reference structure is the height of the user’s nose, and the junk structure is the distance between the outer corners of the user’s eyes. The actual dimensions of the structures remain the same regardless of the imaging distance (i.e.,“Hr” for the reference structure and“Hj” for the junk structure). However, the measured dimensions of the structures at the two different distances can differ. As shown in the columns, the measured dimension at the far distance (e.g., the deterministic structure far dimension taken at one or more far imaging settings, as in stages 808-820 of Figure 8 Figure 8B) is“h1r” for the reference structure and“h1j” for the junk structure, and the measured dimension at the near distance (e.g., the deterministic structure near dimension taken at one or more near distance imaging settings, as in stages 824-836 of Figure 8 Figure 8C) is“h2r” for the reference structure and“h2j” for the junk structure.

[0061] The“Distance” column in Table 1 represents the distance between the imaging system (e.g., the lens of a smartphone camera) and the measured structural feature. The far distance for the reference structure and the junk structure (i.e., corresponding to an imaging system placed according to one or more far imaging settings) are listed in Table 1 as“R1r” and“R1j,” respectively; and the near distance for the reference structure and the junk structure (i.e., corresponding to an imaging system placed according to one or more near distance imaging settings) are listed in Table 1 as“R2r” and“R2j,” respectively. As described herein, some embodiments use relative dimensions. For example, the unit dimension used in Table 1 (“Unit”) is the actual dimension of the reference structure (Hr). The absolute measurement of this dimension can be unknown, but nonetheless all other measurements, dimensions, calculations, etc. can be relative to Hr. In other embodiments, one or more calibrators, and / or other techniques are used to establish an absolute reference dimension. In such embodiments, the reference unit dimension can be any consistent dimension that is independent of the structures of the object being imaged, such as millimeters, the spacing or size of a template (e.g., having evenly spaced tick marks, identical sized geometric shapes, etc.), etc.

[0062] Using the variables defined in Table 1, various relationships can be derived. It can be seen that the imaging distance of the imaging system is mathematically related (e.g., proportional) to the focal length (f) of the imaging system multiplied by the ratio between the actual and measured dimensions of the object: R ~ f * H / h. The value of f can be known to the imaging system (e.g., or any other suitable processing system). In some implementations, f is a constant parameter of the components of the imaging system. In other implementations, f can change with different focal length settings of the imaging system, such as whether the imaging system is used for auto-focus or manual focus. In such other implementations, the imaging system (or any other suitable processing system) can know the current focal setting of the imaging system (e.g., by detecting the relative positions of the lens, aperture, imaging plane, etc.) and can accordingly derive f, such as by looking up the value of f in a lookup table, and / or by calculating f as a function of the relative positions of the imaging components. Furthermore, as mentioned above, Hr can be measured based on its own reference (e.g., Hr is established as a unit dimension, as shown in Table 1), or relative to the calibrator (e.g., as a specific number of millimeters). Thus, given f and Hr, the distance of the reference object (Rr) is inversely proportional to the measured dimension of the reference object hr.

[0063] With respect to another object, various dimensions and other information related to those of the reference object and its own measurements can be derived. The imaging distance of the reference object (Rr) is different from the imaging distance of the other object (Rj) by a certain difference in imaging distance (dj), such that dj = Rr - Rj. For example, assume that one reference dimension extracted at the user’s nose (reference structure) is 300 millimeters from the smartphone lens, while another dimension extracted from the user’s eye (other structure) is 325 millimeters from the smartphone lens; making the other structure 25 millimeters from the imaging system (i.e., dj = 25 millimeters). Furthermore, while the actual dimension of the other structure (Hj) can not be known, it can be assumed that the ratio of Hj to Hr is the same as the ratio of hj to hr (i.e., the measured dimension). The normalization factor of the other structure (Aj) can be defined as hj / hr. Based on the above relationships, the actual dimension of the other structure (Hj) relative to the unit structure dimension (Hr) can be expressed as Hj = Rj * Aj / Rr.

[0064] The actual dimensions of the reference and other structures (Hr and Hj) and the relative difference in imaging distance (dj) between these structures are structural characteristics of the imaged object and do not change with imaging settings. However, the measured dimensions change at different imaging distances, and the normalization factor (Aj) also changes due to the difference in image magnification at different distances. Figures 9A-9DThe changing image magnification in example image 900, acquired by imaging an experimental setup at different imaging distances, is shown. The illustrated experimental setup includes a base having an identical array of points. A platform 910 is placed on the base and also has an array of points identical to each other and to the points on the base. Specifically, all points on the base and platform are circular with identical diameter and spacing, such that the distance between the five points is 17.3 millimeters (mm).

[0065] Each image 900 shows the same experimental setup, but imaged from a different imaging distance. Specifically, the imaging distance between the imaging system and platform 910 was approximately 60 centimeters (cm), 40 cm, 25 cm, and 10 cm for each of images 900a-900d, respectively. Platform 910 was elevated approximately 16.6 mm above the base. Thus, in each image 900, the imaging distance to the base was 16.6 mm greater than the imaging distance to platform 910. Even though the actual dimensions of points on the base and platform 910 were the same, the measured dimensions were different due to the respective differences in imaging distances between platform 910 and the base. Furthermore, due to differences in image magnification, the respective differences were different at different imaging distances, as can be seen by comparing the relative point dimensions between points on the base and points on platform 910 in each image 900.

[0066] For example, in Figure 9A In image 900a, the size of the dashed version of rectangle 920a represents the 17.3 mm span between the five points on platform 910. Placing the solid version of rectangle 920a, which is the same size, against the background of the five points on the base, it can be seen that the measured dimensions of these points (16.6 mm from the imaging system) are slightly smaller than those of the points on platform 910. In contrast, Figure 9D In image 900d, the dashed version of rectangle 920d is again resized to represent the 17.3 mm span between the five points on platform 910. Similarly, a solid version of rectangle 920d of the same size is placed in the background of the five points on the base. Figure 9A On the contrary, Figure 9D As can be seen from the figure, the measured dimensions of other points on the base are significantly smaller than those of the points on the platform 910.

[0067] from Figures 9A-9DAs can be seen, the same 17.3 mm imaging distance difference effectively magnifies the impact on the measured dimension as the object approaches the imaging system. Returning to Table 1, this magnification factor shows the difference between a far normalization factor (A1j) defined as h1j / h1r and a near normalization factor (A2j) defined as h2j / h2r. These normalization factors can be computed from the deterministic structure far dimension in the far imaging setup and the deterministic structure near dimension in the near imaging setup. Based on the measured (or computed) A1j, A2j, h1r, h2r, and image lens focal length f, dj can be tested as:

[0068]

[0069] This relationship established from two (or more) imaging setups can effectively yield a measured dimension (hdj) that should be taken from any particular other structure (j) at any distance (d).

[0070] While Figures 9A-9D The magnification factor shown can help establish three-dimensional information, which also suggests that depth resolution can be poor at larger imaging distances. In some embodiments, one or more imaging setups include a magnifying imaging system. As one example, the inventors have experimentally demonstrated that the passive sensing approach described herein can achieve a depth resolution of approximately 2 mm at a 25 cm distance (e.g., a measured dimension between a point on the pedestal and a point on the platform 910 differs by 8 pixels, and the imaging distance differs by 16.6 mm) using an 8 megapixel camera if the reference dimension is known. Setting the imaging system to magnify by a factor of 2 has also been shown to yield a depth resolution of approximately 1 mm (i.e., approximately twice the depth resolution).

[0071] Figure 10 A flowchart illustrating an illustrative method 1000 of authenticating an authentication object by passively imaging a three-dimensional deterministic structure of the authentication object in accordance with various embodiments herein is shown. As described above, in some implementations, the three-dimensional deterministic structure of the authentication object can be deterministic macro- and / or micro-structures of a human user’s face. It can be assumed that the authentication object being authenticated in the method 1000 has previously gone through a registration process, such as by Figure 8800 in the method 800. At stage 1004, the registration routine is activated. For example, the face recognition authentication routine is activated by an purportedly authorized user of the mobile device in connection with the user's attempt to unlock the phone (from a locked state), authorize a mobile payment transaction, authorize user credentials for an application, and the like. As described herein, the authentication routines can be used for different levels and / or types of authentication. Some embodiments provide biometric authentication, for example, by obtaining sufficiently deterministic structural dimensions of the authentication object to support finding a statistical match with previously registered structural dimensions of the authentication object. Other embodiments additionally, or alternatively, provide spoofing detection, for example, by obtaining three-dimensional dimensions, texture features, and / or other types of information that are often impractical (or practically impossible) to reproduce using spoofing techniques. Other embodiments additionally, or alternatively, provide biometric identification detection, for example, by obtaining sufficiently deterministic structural dimensions of the authentication object to support identifying the authentication object from a large number of candidate authentication objects with sufficient statistical confidence.

[0072] For any biometric authentication, spoof detection, biometric identification and / or other purposes, the registration routine (e.g. Figure 8 Aspects of the authentication routine 1000 (e.g., the enrollment routine in

[00105] ) and the authentication routine 1000 can be used to collect and process different numbers and types of dimensions and dimensional relationships to satisfy potentially competing design considerations. For example, increasing the number of data points relied upon for authentication and / or increasing the threshold confidence level required for a positive authentication can increase the security provided by the authentication routine. However, such increased security may also require more processing time and other processing resources for each authentication determination, may result in more "false positives" (e.g., an authorized user is denied authorization by the routine because the routine cannot achieve sufficient confidence in its determination), and / or may result in frustration in use. Therefore, for certain types of applications (e.g., unlocking a consumer smartphone for typical everyday use), it is desirable to adjust the routine to favor convenience over security to some extent. For example, reducing the number of data points and / or confidence levels used for authentication may reduce the level of security (e.g., generating more false positives, etc.), but may still provide a sufficient level of security for the application while also improving the end-user experience (e.g., by increasing responsiveness, reducing false positives, etc.).

[0073] At stage 1008, embodiments can capture a set of images of an authentication object (e.g., a user's face) using an imaging system. For example, the authentication object is a user's face and the imaging system is integrated into a smartphone, a computer (e.g., a laptop, a desktop, a tablet, etc.), an e-reader device, an automated teller machine, a wearable device, a biometric physical access control system (e.g., a biometric door lock), etc. In some embodiments, the user is prompted and / or otherwise directed to a desired imaging setup. The user can be directed to ensure that the authentication object is at an appropriate imaging distance from the imaging system, to ensure that the authentication object is oriented in a desired manner relative to the imaging system, etc. For example, the user is prompted to look directly at a particular point on a display, thereby tending to direct the user's face (as the authentication object) to a relatively small range of distances and directions that are likely to match a registered set of imaging conditions.

[0074] During the capture at stage 1008, the imaging system is at an authentication imaging distance from the first deterministic structure that is remote from the authentication object. As described herein, the imaging system is at some respective imaging distances from each of one or more other deterministic structures. Each of these respective imaging distances is a respective delta (difference) from the authentication imaging distance. The first deterministic structure and the one or more other deterministic structures are visible in the set of images (e.g., fewer than all of the deterministic structures can be seen in any set of images). The "imaging distance" can be referenced to any suitable component of the imaging system, such as a focal plane of a lens of the imaging system. In some implementations, the imaging distance expected for authentication is closer than the "far distance" used for enrollment and / or closer than the "near distance" used for enrollment. For example, in the context of facial recognition, a user can tend to hold a smartphone at a comfortable distance from his or her face; this can naturally be between the distances used to capture images during enrollment. In other implementations, any other imaging distance can be accommodated, so long as the distance is far enough from the imaging system that each captured image includes a sufficient set of deterministic structures to support providing a desired authentication level of the authentication object, and so long as the distance is close enough to the imaging system to provide sufficient depth resolution to support providing a desired authentication level of the authentication object. As described above, some embodiments support manually and / or automatically zooming the imaging system to enhance depth resolution. For example, embodiments can automatically zoom or prompt the user to manually zoom based on detecting that the imaging distance exceeds a threshold imaging distance from the authentication object.

[0075] At stage 1012, embodiments can process the set of images to measure at least a first authentication dimension of the first deterministic structure and a second authentication dimension of the second deterministic structure. At least the first and second deterministic structures are visible in the set of images. For example, as referenced above, the first authentication dimension can be a distance between the first deterministic structure and the second deterministic structure, and the second authentication dimension can be a distance between the second deterministic structure and a third deterministic structure. In some embodiments, the set of images is processed to measure the first authentication dimension and the second authentication dimension in a single pass through the set of images. In other embodiments, the set of images is processed to measure the first authentication dimension and the second authentication dimension in two or more passes through the set of images. Figure 8The first deterministic structure can be any suitable reference structure from which a first authentication measurement (e.g., a height of a nose) is obtained as a deterministic reference dimension. Each additional deterministic structure can be any suitable macrostructure or microstructure from which a respective additional authentication measurement is obtained as another respective deterministic structure dimension (e.g., a width of a left eye, a dimension within a texture feature along a user's forehead, etc.).

[0076] At stage 1016, embodiments can obtain the enrollment dimensions from the enrollment database obtained from measuring at least the first deterministic structure and the second deterministic structure during a prior enrollment process. In some embodiments, the obtaining at stage 1016 includes obtaining a first enrollment dimension obtained from measuring the first deterministic structure during the enrollment process, and a second enrollment dimension obtained from measuring the second deterministic structure during the enrollment process, and an enrollment imaging distance delta (e.g., referred to above as "dj") computed between the first and second deterministic structures during the enrollment process. The performance of stage 1016 can assume that the particular first authentication measurement and the respective additional authentication measurement generated at stage 1012 correspond to structure dimensions previously enrolled during an enrollment process. For example, with reference to Table 1 above, the first enrollment dimension can be Hr, the second enrollment dimension can be Hj, and the enrollment imaging distance delta can be dj. The measurements obtained at stage 1016 can be obtained from any suitable enrollment database or databases. For example, as described with reference to Table 1 above, the relevant enrollment information can be stored in a single enrollment database, a near-far enrollment database, etc. Figure 8

[0077] ​At stage 1020, embodiments can calculate an authentication imaging distance based on a relationship between the first authentication dimension and the enrollment dimension, and can calculate an expected imaging distance for the second deterministic structure based on a relationship between the authentication imaging distance and the enrollment dimension; in some embodiments, the authentication imaging distance is calculated based on the first enrollment dimension, the first authentication dimension, and a focal point characteristic of the imaging system. As described above, the first imaging distance (R3r) can be functionally related to the focal length (f) of the imaging system multiplied by the ratio between the first enrollment dimension (Hr) and the first authentication dimension (h3r). For example, with reference to the equation provided above, the first imaging distance can be calculated as: R3r = f*Hr / h3r. In some embodiments, the expected imaging distance is calculated for the second deterministic structure based on the first imaging distance and an enrollment imaging distance increment. For example, during enrollment, the enrollment imaging distance increment is calculated based on a relationship between the deterministic structure far dimension and the deterministic structure near dimension. Notably, the dimensions (Hr, Hj, and dj) acquired in stage 1016 are actual structure dimensions of the authentication object that are independent of any imaging setup (e.g., the width of one eye does not change based on the position of the imaging system, focal length, illumination, etc.). Thus, if the authentication object is the same as the enrollment object, Rj should always be farther away from Rr regardless of the value of Rr. Accordingly, at stage 1020, the expected imaging distance (R3je) for the second deterministic structure is greater than (or less than, if dj is negative) the first imaging distance (R3r) also calculated in stage 1020.

[0078] At stage 1024, embodiments can calculate an authentication normalization factor as a ratio between the first and second authentication dimensions, and can calculate an expected normalization factor for the second deterministic structure at the expected imaging distance based on the enrollment dimension and the authentication imaging distance. As described above, the normalization factor (Aj) for the second deterministic structure at a particular set of imaging distances for authentication represents the ratio of the measured dimensions of the deterministic structure at that imaging distance. Thus, consistent with the equation provided above, the authentication normalization factor (A3j) can be calculated as a ratio between the first and second authentication dimensions: A3j = h3j / h3r. As derived with reference to Table 1, the expected normalization factor for the second deterministic structure at any particular imaging distance is functionally related to the actual size of the second deterministic structure and the ratio of the imaging distances between the first and second deterministic structures. Based on the enrollment information and with reference to the equation provided above, the expected normalization factor (A3je) for the particular set of imaging distances for authentication can be calculated as: A3je = Hj*R3r / R3je.

[0079] At stage 1028, embodiments can determine whether to authenticate the authentication object based on comparing the authentication normalization factor to the expected normalization factor. The expected normalization factor is computed based on data acquired by a "true" authentication object during a previous registration process, and the authentication normalization factor is computed based on data measured by the "claimed" (i.e., the authentication object claims to be a true authentication object and authenticates accordingly) authentication object during the current authentication (at stage 1012). If the claimed authentication object is not a true authentication object, then the authentication normalization factor can not match the expected normalization factor, and the authentication can be denied. If the claimed authentication object is indeed a true authentication object, then the authentication normalization factor will likely match the expected normalization factor, and the authentication can be granted.

[0080] The match can be determined in any suitable manner, and to any predetermined level of confidence. In some embodiments, the determination of a match involves determining a match between at least a predetermined minimum number of certainty structure dimensions. A single iteration of the method 1000 can provide only a single data point for determining a match. For example, the first and second certainty structures are used together (one as a reference for the other) to establish a single authentication normalization factor for comparison to a single expected normalization factor. Embodiments can iterate through the method 1000 (e.g., through stages 1012-1028) multiple times to acquire multiple authentication normalization factors and expected normalization factors for the same second certainty structure and / or one or more additional certainty structures; and determining whether there is a match at stage 1032 can be based on matching some or all of the multiple authentication normalization factors to some or all of the expected normalization factors.

[0081] In some embodiments, determining whether there is a match at stage 1032 can involve additional computations and / or processing. For example, since the enrollment and authentication images are captured at different times, it can be expected that the set of images, measured dimensions, dimensional relationships, etc. can not exactly match. Thus, the determination of a match can involve applying statistical computations to determine a confidence between the enrollment and authentication data sets. For example, a "match" can be defined as satisfying a threshold magnitude (e.g., 95%) of statistical correlation between at least one set or multiple sets of relevant dimensions. In some cases, one or more types of three-dimensional transformations (e.g., rotation, tilt, scaling, etc.) are applied to attempt to establish an equivalent (or substantially equivalent) dimensional basis for performing the determination of a match between the enrollment and authentication data sets, prior to performing any type of match determination. In some embodiments, a settings portal can be provided through which an authorized user, administrator, provider, etc. can adjust parameters such as a threshold statistical confidence for determining a match, a number of data points to collect and use, etc. Adjusting such settings can help change an operational balance between a user experience and a level of security. For example, an authorized user can adjust the setting to increase a level of security at the expense of increased processing resources and potentially increased false positives (and associated frustration of false denied access); or, alternatively, an authorized user can adjust the setting to increase a level of convenience (e.g., with faster response times, fewer false positives, etc.) at the expense of reduced security (e.g., potentially increased false positives).

[0082] In some embodiments, at stage 1036, a user-perceptible authentication result can be output based on the determination at stage 1032 to indicate whether the authentication result is an authorization or a denial of authentication. For example, a visible, audible, tactile, and / or other response can be output by the imaging system, or by any suitable system communicatively coupled with the imaging system. In one example, in the context of facial recognition for smartphone access control, a successful authentication determination at stage 1032 can result in the smartphone being unlocked at stage 1036 to allow access (e.g., the output at stage 1036 manifests as a display switching from a lock screen). In another example, in the context of a biometric physical access control system, a successful authentication determination at stage 1032 can result in a physical door being unlocked at stage 1036 to allow access (e.g., the output at stage 1036 is a signal to a door lock mechanism). In another example, in the context of general biometric authentication, a successful authentication determination at stage 1032 can result in an output at stage 1036 in the form of a light being lit, a graphical display element on a display changing appearance, a component vibrating, an audible chime being played through a speaker, etc.

[0083] Figure 11A flowchart showing an example gating process 1100 for access control, in accordance with various embodiments, is shown. For example, a user can seek access to a smartphone having an integrated image sensor system, such as described herein. Access to the smartphone is locked until the user successfully passes a user facial biometric verification based on previously registered data. Biometric verification (or authentication) generally refers to verifying a biometric of a candidate user (or other object) against a corresponding biometric of a previously registered user. For example, even though the term "biometric" relates to a characteristic of a living being, the term is more generally used herein to refer to a structure having a dimension of certainty that is sufficiently unique to the object being authenticated to be useful for authentication. For example, while the biometric structure of a person's face can be part of the biology of the person, inorganic and / or non-living objects can include structures having a sufficient dimension of certainty and uniqueness to serve as a characteristic of that particular object.

[0084] Biometric verification can be much simpler than so-called biometric identification. For example, biometric identification can seek to determine the identity of a candidate user from a general population of users, such as determining whether a fingerprint matches any one in a large database of fingerprints to at least some threshold level of confidence; whereas biometric verification can start with a set of hypothesized pre-registered users (e.g., one, or a relatively small number), and can seek to determine whether a current candidate user appears to match one of the hypothesized set of users to a threshold level of confidence. Biometric access control systems, such as those in example smartphones, are generally based on biometric verification. For example, a smartphone (or similar identification card, electronic door lock, etc.) can be associated with only a single authorized user, and the function of the system is to determine whether a candidate user attempting access (e.g., statistically) is the authorized user. Such a function does not require the system to search a large database in an attempt to identify the candidate user.

[0085] At a pre-biometric trigger stage 1110, embodiments can wait for detection of a candidate image or images, which can trigger further biometric verification. In some cases, the pre-biometric trigger stage 1110 is itself triggered to begin by one or more other trigger conditions. For example, the pre-biometric trigger stage 1110 begins after a user presses a button, enters a code, opens a device, etc. In some other cases, an image sensor can continuously, periodically, or otherwise acquire images. The images can be dynamically processed to detect a set of image data that is characteristic of a human face, or of a candidate for biometric verification. For example, detection of certain trackable structures (e.g., macro and / or micro structures) in a particular pattern (e.g., at a relative position, size, etc.) indicates to the system that the captured image is a candidate face image for biometric processing. In some implementations, this stage 1110 can use various techniques to improve detection of such trackable structures. For example, stage 1110 can include focusing an imaging system based on one or more parameters, such as based on a chromatic component; and / or 1110 can include analyzing individual chromatic components of raw image data (e.g., including statistical analysis including computation of an image intensity map, etc.); and / or stage 1110 can include correcting parameters of the imaging data, such as contrast, spectral reflectance, spectral illumination non-uniformity, surface transmittance, etc.

[0086] At a biometric verification stage 1120, the same and / or different trackable structures are used for biometric verification of a pre-registered user. In some implementations, the imaging data acquired in stage 1110 is sufficient for biometric verification in stage 1120. In other implementations, additional and / or different imaging data is acquired, such as high-definition data with multiple chromatic components. In some embodiments, stage 1120 can include adjusting and / or re-orienting acquired data, and / or correcting the size and / or orientation of the data. For example, as described above, certain trackable structures have known dimensions, certain distances between trackable structures are known, etc. Comparing such known information with acquired information can provide information about the distance between the imaged object and the imaging system (e.g., an object appears smaller when it is farther away from the imaging system), and / or the orientation of the imaged object relative to the imaging system (e.g., when an imaged object is tilted, a set of its trackable structures tilt in a deterministic manner). In some implementations, parameters of the imaging system are also known and can be used in this stage 1120.

[0087] Some embodiments of the biometric verification stage 1120 follow a process similar to that described in U.S. Patent No. 8,245,414, which is incorporated by reference herein in its entirety. Figure 10the stages of method 1000 in FIG. 11. For example, the pre-biometric feature triggering stage 1110 captures relatively low resolution information that is generally intended to quickly inform the imaging system that there appears to be a candidate object of the type used for authentication in the imaging field of the imaging system. At the biometric feature verification stage 1120, the candidate object becomes a purported authenticating object, and the biometric feature verification stage 1120 attempts to determine whether to authenticate the purported authenticating object. Some embodiments end with a successful pass of the biometric feature verification stage 1120. For example, the output of a verification signal is triggered by the biometric feature verification stage 1120 (e.g., in the form of a Figure 10 In some embodiments, the verification signal (or an additional signal) triggers the access control system to grant access (e.g., a smartphone to unlock) at stage 1140.

[0088] Other embodiments also include a spoof detection stage 1130. For example, a successful pass of the biometric feature verification at stage 1120 can trigger an additional hurdle of the spoof detection stage 1130 that must also be passed before the access control system is allowed access at stage 1140. As described above, such a spoof detection stage 1130 can use information acquired in the biometric feature verification stage 1120, and / or can acquire any suitable information to determine whether the candidate object is a spoof. For example, establishing even one, or a small number of three-dimensional relationships between macro-structures of the purported authenticating object can be sufficient to distinguish a “real” three-dimensional human face from a printed photograph of the same human face. However, these same measurements can not be sufficient to detect three-dimensional spoofing, such as a 3D sculpture of a wax or latex human face, a mask prosthetic worn on a human face, a high-resolution photograph print, etc. In such cases, micro-structural dimensions, textural features, multi-spectral information, and / or other information can be used in the spoof detection stage 1130 to detect other types of spoofing (and can be acquired in the spoof detection stage 1130 if not previously acquired).

[0089] Some embodiments can include only one or two stages of the flowchart 1100, and the various stages can be performed in any order. In some embodiments, the spoof detection stage 1130 and the biometric verification stage 1120 are performed sequentially. For example, the biometric verification stage 1120 is successful triggers the start of the spoof detection stage 1130. In other embodiments, the biometric verification stage 1120 and the spoof detection stage 1130 are performed simultaneously (i.e., at least partially in parallel). In some embodiments, some or all of the stages can be triggered independently. For example, a user can explicitly trigger the biometric verification stage 1120 such that stage 1120 is not responsive to successful identification of a candidate in stage 1110. Similarly, a user can explicitly trigger the spoof detection stage 1130 without triggering the associated biometric verification stage 1120. For example, there can be a situation in which a user desires to know whether an object is spoofed, but is unable to determine any type of biometric verification of the object.

[0090] Figure 12 A diagram of one embodiment of a computer system 1200 that can implement various system components and / or perform various steps of the methods provided by various embodiments is provided. It should be noted that, Figure 12 Any or all of the components can be used in Figure 12 The various system elements can be implemented in a relatively

[0091] The illustrated computer system 1200 includes hardware elements (or, in appropriate contexts, software elements) that can be electrically coupled via a bus 1205 (or can otherwise be in communication with one another). The hardware elements can include one or more processors 1210, including without limitation one or more general-purpose processors and / or one or more special-purpose processors (such as digital signal processing chips, graphics acceleration processing units, video decoders, and / or other like processors). The processor(s) 1210 can be configured to implement a Figure 2 The illustrated control and processing module 230. Some embodiments include one or more input / output (I / O) devices 1215. In some implementations, the I / O devices 1215 include human-perceptible interface devices, such as buttons, switches, keypads, indicators, displays, and / or the like. In other implementations, the I / O devices 1215 include circuit-level devices, such as pins, dip switches, and / or the like. In some implementations, the computer system 1200 is a server computer for interfacing with additional computers and / or devices, such that the I / O devices 1215 include various physical and / or logical interfaces (e.g., ports, and / or the like) to facilitate hardware-to-hardware coupling, interaction, control, and / or the like. In some embodiments, the I / O devices 1215 implement the output interface 250 of Figure 2

[0092] ​The computer system 1200 can further include (and / or be in communication with) one or more non-transitory storage devices 1225, which can include, without limitation, local and / or network accessible storage, and / or can include, without limitation, a disk drive, a drive array, an optical storage device, a solid-state storage device such as a random access memory ("RAM") and / or a read-only memory ("ROM"), which can be programmable, flash- updateable and / or the like. Such storage devices can be configured to implement any appropriate data stores, including without limitation, various file systems, database structures, and / or the like. In some embodiments, the storage devices 1225 include a database 240 in which the registration database 240 is stored. Figure 2

[0093] The computer system 1200 can further include (and / or be in communication with) one or more non-transitory storage devices 1225, which can include, without limitation, local and / or network accessible storage, and / or can include, without limitation, a disk drive, a drive array, an optical storage device, a solid-state storage device such as a random access memory ("RAM") and / or a read-only memory ("ROM"), which can be programmable, flash- updateable and / or the like. Such storage devices can be configured to implement any appropriate data stores, including without limitation, various file systems, database structures, and / or the like. In some embodiments, the storage devices 1225 include a database 240 in which the registration database 240 is stored. Figure 2

[0094] ​​Embodiments of computer system 1200 can also include working memory 1235, which can include RAM or ROM devices, as discussed above. Computer system 1200 can also include software elements, shown as being currently located within working memory 1235, including an operating system 1240, device drivers, executable libraries, and / or other code such as one or more applications 1245, which can implement methodologies provided by various embodiments, and / or can be designed to implement methods and / or configure systems provided by other embodiments, as discussed herein. Merely by way of example, one or more procedures as described with respect to the method implementations discussed herein can be implemented as code and / or instructions executable by a computer (and / or a processor within a computer); in an aspect, then, such code and / or instructions can be used to configure and / or adapt a general purpose computer (or other device) to perform one or more operations in accordance with the described methods. A set of these instructions and / or code might be stored on a non-transitory computer-readable storage medium, such as the non-transitory storage device(s) 1225 described above. In some cases, the storage medium might be incorporated within a computer system, such as computer system 1200. In other embodiments, the storage medium might be separate and / or removable from the computer system (e.g., a USB drive, an optical disc, etc.), and / or provided in an installation package, such that the storage medium can be used to program, configure, and / or adapt a general purpose computer with the instructions / code stored thereon. These instructions might take the form of executable code, which is executable by the computer system 1200 and / or might take the form of source code, which can be compiled and / or interpreted to provide executable code.

[0095] It will be apparent to those skilled in the art that substantial variations can be made in accordance with specific requirements. For example, customized hardware might also be used, and / or particular elements might be implemented in hardware, software (including portable software, such as applets, etc.), or both. Further, connection to other computing devices such as network input / output devices can be employed.

[0096] As described above, in one aspect, some embodiments can employ a computer system (such as the computer system 1200) to perform portions or all of the process for implementing various embodiments of the application. In accordance with a set of embodiments, the computer system 1200 performs some or all of the processes for various embodiments of the application in response to the processor 1210 executing one or more sequences of one or more instructions (which might be incorporated into the operating system 1240 and / or other code, such as an application program 1245) contained in the working memory 1235. Such instructions can be read into the working memory 1235 from another computer readable medium, such as one or more of the non-transitory storage device(s) 1225. The execution of the sequences of instructions contained in the working memory 1235 might cause the processor 1210 to perform the processes described herein.

[0097] As used herein, the terms "machine-readable medium," "computer-readable storage medium" and "computer-readable medium" refer to any medium, non-transitory, that participates in providing data that causes a machine to operate in a specific fashion. Such a medium might take many forms, in implementations where the computer system 1200 is implemented. Various computer-readable media might be involved in storing and / or carrying the sequences of instructions to the processor 1210 for execution. In many implementations, a computer-readable medium is a physical and / or tangible storage medium. Such a medium can take many forms, including but not limited to, non-volatile media, and / or volatile media. Non-volatile media includes, for example, optical or magnetic disks, such as the non-transitory storage device(s) 1225. Volatile media includes, for example, dynamic memory, such as the working memory 1235.

[0098] Common forms of physical and / or tangible computer-readable media include, for example, a floppy disk, a flexible disk, a hard disk, magnetic tape, or any other magnetic medium, a CD-ROM, any other optical medium, any other physical medium with patterns of holes, a RAM, a PROM, EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave as described hereinafter, or any other medium from which a computer can read instructions and / or code. Various forms of computer-readable media can be involved in carrying one or more sequences of one or more instructions to the processor 1210 for execution. Merely by way of example, the instructions can initially be carried on a magnetic disk of a remote computer. A remote computer might load the instructions into its dynamic memory and send the instructions as signals over a transmission medium to be received and / or executed by the computer system 1200.

[0099] It should be understood that when an element or component is referred to as being "connected to" or "coupled to" another element or component, it can be directly connected or coupled to the other element or component or intervening elements or components can be present. In contrast, when an element or component is referred to as being "directly connected to" or "directly coupled to" another element or component, there are no intervening elements or components present. It will be appreciated that, although terms such as "first" and "second" can be used herein to describe various elements or components, these elements or components should not be limited by these terms. These terms are only used to distinguish one element or component from another. Thus, a first element, component, or step discussed below could be termed a second element, component or step without departing from the teachings of the present application. As used herein, the terms "logic low," "low state," "low level," "logic low level," "low," or "0" are used interchangeably. The terms "logic high," "high state," "high level," "logic high level," "high," or "1" are used interchangeably.

[0100] As used herein, the terms "a," "an," and "the" can include both singular and plural referents. It will be further understood that the terms "comprises," "comprising," "includes," "including," "has," "having," "has" and "having," when used in this specification, specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof. Conversely, the term "comprising" in this specification indicates the inclusion of the stated features, steps, operations, elements, and / or components, but not the exclusion of additional features, steps, operations, elements, and / or components. Furthermore, as used in this specification, the word "and / or" can mean and, but it can also mean and / or, one or the other, and / or both.

[0101] Although the present application is described herein with reference to exemplary embodiments, the description is not intended to be construed in a limiting sense. Rather, it is the intent that the scope of the application be defined by the claims appended hereto. Exemplary embodiments are described herein with reference to the accompanying drawings, which are meant to be exemplary and illustrative, and not limiting the scope of the application. Various modifications and changes can be made with respect to the exemplary embodiments described herein, and it is intended to claim all such modifications and changes as fall within the scope of the application. Accordingly, the drawings and descriptions are to be regarded as illustrative in nature, and not restrictive.

[0102] Moreover, some of the features of the preferred embodiments of this application can be used to advantage without the corresponding use of other features. As such, the foregoing description should be considered as merely illustrative of the principles of the application, and not in limitation thereof. Since this application is of a generic type, many modifications and variations of the specific implementation described can be constructed without departing from the spirit or scope of the application. Accordingly, it should be understood that the application has been described by way of illustration only and not limitation. It is the intention that all such modifications and changes be considered as being within this application.

Claims

1. A method for authenticating an authentication object based on passive imaging, the method comprising: capturing a set of images of the authentication object using an imaging system located at an authentication imaging distance from a first deterministic structure of a plurality of deterministic structures of the authentication object visible in the set of images; processing the set of images to measure at least a first authenticated dimension of the first deterministic structure and a second authenticated dimension of a second deterministic structure of the plurality of deterministic structures; Obtaining a registration dimension from a registration database, the registration database being obtained by measuring at least the first deterministic structure and the second deterministic structure in a previous registration process; calculating the authentication imaging distance based on a relationship between the first authentication dimension and the registration dimension, and calculating an expected imaging distance for the second deterministic structure based on a relationship between the authentication imaging distance and the registration dimension; calculating an authentication normalization factor as a ratio between the first authentication dimension and the second authentication dimension, and calculating an expected normalization factor for the second deterministic structure at the expected imaging distance based on the registration dimension and the authentication imaging distance; as well as The authentication subject is authenticated based on comparing the authentication normalization factor with the expected normalization factor.

2. The method according to claim 1, further comprising: Outputting a user-perceivable authentication result to indicate whether the authentication result is authorization or rejection of authentication.

3. The method according to claim 1, wherein The obtaining of the registration dimension includes obtaining, from the registration database, a first registration dimension obtained by measuring the first deterministic structure in the previous registration process, a second registration dimension obtained by measuring the second deterministic structure in the previous registration process, and a registration imaging distance increment between the first deterministic structure and the second deterministic structure calculated in the registration process.

4. The method according to claim 3, wherein: The second registration dimension, the first authentication dimension, and the second authentication dimension are measured according to unit dimensions defined in the previous registration process based on the first registration dimension.

5. The method according to claim 1, wherein The authentication imaging distance is calculated as a function of a focus characteristic of the imaging system and a ratio between a first registration dimension and the first authentication dimension, the first registration dimension being obtained by measuring the first deterministic structure during the previous registration process.

6. The method according to claim 1, wherein An expected imaging distance of the second deterministic structure is calculated by adding the certified imaging distance to a registration imaging distance increment, the registration imaging distance increment being a difference between respective imaging distances from the imaging system to each of the first and second deterministic structures during the previous registration process.

7. The method according to claim 1, wherein An expected normalization factor of the second deterministic structure is calculated as a function of a second registration dimension obtained by measuring the second deterministic structure in the previous registration process and a ratio between the certified imaging distance and the expected imaging distance.

8. The method according to claim 1, wherein The capturing includes capturing at least one of the set of images on the authentication object with the imaging system magnified to form at least one of the set of images with increased depth resolution.

9. The method according to claim 1, wherein: The plurality of deterministic structures includes the first deterministic structure, and a plurality of other deterministic structures including the second deterministic structure; Processing the set of images includes processing the set of images to measure the first authentication dimension of the first deterministic structure and respective other authentication dimensions of each of the plurality of other deterministic structures; The acquiring of the registration dimension comprises acquiring the registration dimension obtained by measuring the reference deterministic structure and the plurality of other deterministic structures during the previous registration process; The calculating the expected imaging distance includes calculating a respective expected imaging distance for each of the plurality of other deterministic structures; said calculating an authentication normalization factor comprises calculating a respective authentication normalization factor for each of said plurality of other deterministic structures; The calculating the expected normalization factor includes calculating a respective expected normalization factor for each of the plurality of other deterministic structures at a respective expected imaging distance; as well as The authentication object is authenticated based on comparing a plurality of respective authentication normalization factors to a corresponding plurality of respective expected normalization factors.

10. The method according to claim 9, wherein: At least a first of the plurality of other deterministic structures of the authentication object is a macrostructure; and At least a second one of the plurality of other deterministic structures of the authentication object is a microstructure.

11. The method according to claim 1, wherein The authenticating the authentication object includes biometric verification and spoofing detection based on at least a portion of a plurality of deterministic structures of the authentication object.

12. A system for authenticating an object based on passive imaging, the system comprising: a registration database storing a plurality of registration dimensions obtained by measuring a plurality of deterministic structures of an authentication object in a previous registration process; an imaging subsystem for capturing a set of images of the authentication object, wherein the authentication object is an authentication imaging distance from a first deterministic structure of the plurality of deterministic structures; as well as A control and processing module having one or more processors and a memory having a set of instructions stored thereon, which when executed cause the one or more processors to perform the following steps: processing the set of images to measure at least a first authenticated dimension of the first deterministic structure and a second authenticated dimension of a second deterministic structure of the plurality of deterministic structures; calculating the authentication imaging distance based on a relationship between the first authentication dimension and the registration dimension; calculating an expected imaging distance for the second deterministic structure based on a relationship between the authenticated imaging distance and the registered dimension; calculating an authentication normalization factor as a ratio between the first authentication dimension and the second authentication dimension; calculating an expected normalization factor for the second deterministic structure at the expected imaging distance based on the registration dimension and the certified imaging distance; as well as The authentication object is authenticated based on comparing the authentication normalization factor with the expected normalization factor.

13. The system of claim 12, further comprising: The output interface communicates with the control and processing module to respond to and instruct the control and processing module to authenticate the authentication object and output a user-perceivable authentication result.

14. The system of claim 12, further comprising: an electronic access control system having a registration database, said imaging subsystem and said control and processing module integrated therein, The electronic access control system is used to authorize or deny access to physical or electronic resources based on the authentication.

15. The system of claim 12, wherein: The plurality of registration dimensions include a first registration dimension obtained by measuring the first deterministic structure during the previous registration process; as well as The first authentication dimension and the second authentication dimension are measured according to unit dimensions defined in the previous registration process based on the first registration dimension.

16. The system of claim 12, wherein: The certified imaging distance is calculated as a function of a focusing characteristic of the imaging subsystem and a ratio between a first registration dimension and the first certified dimension, the first registration dimension being one of the plurality of registration dimensions obtained by measuring the first deterministic structure during the prior registration process.

17. The system of claim 12, wherein: An expected imaging distance is calculated for the second deterministic structure by adding the certified imaging distance to a registered imaging distance increment representing a difference between respective imaging distances from the imaging subsystem to each of the first and second deterministic structures during the previous registration process.

18. The system of claim 12, wherein: An expected normalization factor for the second deterministic structure is calculated as a function of a second registration dimension, one of the plurality of registration dimensions obtained by measuring the second deterministic structure in the previous registration process, and a ratio between the certified imaging distance and the expected imaging distance.

19. The system of claim 12, wherein: The plurality of deterministic structures includes the first deterministic structure, and a plurality of other deterministic structures including the second deterministic structure; Processing the set of images includes processing the set of images to measure a first authentication dimension of the first deterministic structure and a respective other authentication dimension of each of the plurality of other deterministic structures; The registration dimension is a registration dimension obtained by measuring the reference deterministic structure and the plurality of other deterministic structures during the previous registration process; The calculating the expected imaging distance includes calculating a respective expected imaging distance for each of the plurality of other deterministic structures; said calculating an authentication normalization factor comprises calculating a respective authentication normalization factor for each of said plurality of other deterministic structures; The calculating the expected normalization factor includes calculating a respective expected normalization factor for each of the plurality of other deterministic structures at a respective expected imaging distance; as well as The authentication object is authenticated based on comparing a plurality of respective authentication normalization factors to a corresponding plurality of respective expected normalization factors.

20. The system of claim 19, wherein: At least a first of the plurality of other deterministic structures of the authentication object is a macrostructure; and At least a second one of the plurality of other deterministic structures of the authentication object is a microstructure.

Images