BR112021002904A2 - methods and apparatus for facial recognition - Google Patents

Abstract

METHODS AND APPARATUS FOR FACIAL RECOGNITION. Aspects of the present disclosure include methods for generating a heat map including a plurality of points samples having a plurality of characteristic values associated with the detected non-visible light, identifying one or more macroblocks, each including a subset of the plurality of sampling points, calculating a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks, generating a first matrix including a plurality of weighted values by calculating the plurality of weighted values based on the number of occurrences of the local standard value and corresponding sizes of one or more macroblocks, assigning an index unique to each of the plurality of weighted values, generating a second matrix of the unique index ranking the plurality of weighted values, and generating a third matrix including a plurality of ranking distances.

Description

METHODS AND APPARATUS FOR FACIAL RECOGNITION CROSS REFERENCE TO RELATED ORDERS

[001] This application is a continuation in part of United States patent application No.

15/649,144, filed July 13, 2017, which is a continuation of United States Patent Application No. 14/022,080, filed September 9, 2013, now United States Patent No.

9,740,917, issued August 22, 2017, which claims the benefit of United States Provisional Application No. 61/792,922, filed March 15, 2013, and United States Provisional Application No.

61/698,347, filed September 7, 2012, the contents of which are incorporated by reference in their entirety.

BACKGROUND

[002] There has been a growing need for stronger identity verification to protect personal assets, both physical and electronic. For example, it is important to control access to facilities, vehicles and personal property so that only authorized applicants are allowed access. A requester can be a user/person requesting access to controlled access assets and/or infrastructure. In a traditional example, a requester might load and use a key,

which is designed to fit into a lock to allow the key seeker to open the lock and gain entry. A loss or damage to the key, however, can make access impossible. In another example, an applicant can use a keyring to remotely lock or unlock a vehicle's doors, for example, by pressing a button on the keyring to generate an infrared ("IR") or radio frequency ("RF") signal, which is detected by a sensor in the vehicle, which controls the doors. These vehicle keyless access systems may require the applicant to operate the ignition system. Other similar keyless access implementations may involve inserting and presenting a swipe or similar card into an opening or a card reader/detector, or allowing an authorized requester to enter a numeric or alphanumeric code on a provided keyboard.

In each of these conventional techniques, however, it is very difficult to determine whether the person holding the key/card is the authorized requester. An imposter can steal or duplicate a valid key and gain unauthorized access to location, vehicle and/or personal property.

[003] Although traditional biometric access control systems can mitigate some shortcomings of key/card based access control systems, there may be limitations as well. Traditional biometric sensors, such as iris detection sensors, can be limited to specific light conditions, significantly reducing the effectiveness of the biometric sensor as well as the possible environments for applying it.

The performance of biometric sensors can be compromised in direct sunlight due to glare, shadows and other artifacts. Even with the advent of mega-pixel camera technology, the characteristics of each face can be obscured by ambient lighting, the position of the face, changes in the face, the background behind the face and the quality of the camera. Motion blur, insufficient resolution, environmental impacts, lighting, background, and camera angles all collude to obscure subject details, making heterogeneous facial recognition difficult (matching video and other probe images to large databases of frontal photographs ).

[004] Other factors can also increase the false acceptance and/or false recognition rates of traditional biometric sensors. For example, biometric sensors also have difficulty obtaining the necessary data in the absence of light. Light source shading and other changes in intensity can create contrasts in the face that can be misinterpreted as facial features and/or slightly distort the measurement of actual facial features. Another important source of inaccuracy is the increased probability of similar measured features across faces in a growing population. Furthermore, the problem of capturing the characteristics of each face can be aggravated by the desire for low maintenance and/or low complexity face recognition systems. Therefore, improvement in access control may be desired.

SUMMARY

[005] The following is a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not a comprehensive overview of all aspects contemplated and is not intended to identify the key or critical elements of all aspects, nor to delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that will be presented later.

[006] Some aspects of the present disclosure include methods for generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light, identifying one or more macroblocks, each including a subset of the plurality of sampling points, selecting a local default value, calculating a number of occurrences of the local default value within each subset of the plurality of sampling points for each of the one or more macroblocks, generating a first matrix including a plurality of weighted values calculating the plurality of weighted values based on the number of occurrences of the local default value and corresponding sizes of one or more macroblocks, assigning a unique index to each of the plurality of weighted values, generating a second matrix of the unique index ranking the plurality of weighted values and generating a third matrix including a plurality of classification distances.

[007] Certain aspects of the present disclosure include a faceplate having an illumination source configured to emit non-visible incident light, an optical sensor configured to detect a detected non-visible light, wherein the detected non-visible light includes non-visible light. reflected and radiated non-visible light, one or more processors operatively coupled to the light source and optical sensor, one or more processors are configured to build a biometric model of a requester requesting access to an entry point, generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light,

identifying one or more macroblocks, each includes a subset of the plurality of sampling points, selecting a local default value, calculating a number of occurrences of the local default value within each subset of the plurality of sampling points for each of the one or more macroblocks, generating a first array including a plurality of weighted values by calculating the plurality of weighted values based on the number of occurrences of the local standard value and corresponding sizes of one or more macroblocks, assigning a unique index to each of the plurality of weighted values, generating a second matrix of the unique index ranking the plurality of weighted values and generating a third matrix including a plurality of ranking distances.

[008] Aspects of the present disclosure include a computer-readable medium having code stored therein that, when executed by one or more processors, causes one or more processors to execute code to generate a heat map including a plurality of points of sampling having a plurality of characteristic values associated with detected non-visible light, code to identify one or more macroblocks each includes a subset of the plurality of sampling points, code to select a local standard value, code to calculate a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks, code for generating a first matrix including a plurality of weighted values, calculating the plurality of weighted values based on the number of occurrences of the local pattern value and corresponding sizes of one or more macroblocks, code to assign a unique index to each of the plurality of weighted values, generating a second matrix of the unique index ranking the plurality of weighted values, and code to generate a third matrix including a plurality of ranking distances.

[009] An aspect of the present disclosure includes a system having means for generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light, means for identifying one or more macroblocks each including a subset of the plurality of sampling points, means for selecting a local standard value, means for calculating a number of occurrences of the local standard value within each subset of the plurality of sampling points for each of the one or more macroblocks, means for generating a first array including a plurality of weighted values calculating the plurality of weighted values based on the number of occurrences of the local standard value and corresponding sizes of one or more macroblocks, means for assigning a unique index to each of the plurality of weighted values, generating a second matrix of the single index sorting the plurality of weighted values and means for generating a third matrix including a plurality of classification distances.

[0010] Aspects of the present disclosure include an infrastructure having a controlled access entry point, a faceplate configured to emit non-visible incident light on a face of a requester, detect a non-visible light detected on the face of the requester, in that the detected non-visible light includes a reflected non-visible light and a radiated non-visible light, generate a biometric model of the applicant by generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light , identify one or more macroblocks each including a subset of the plurality of sampling points, select a local default value, calculate a number of occurrences of the local default value within each subset of the plurality of sampling points for each of the one or more macroblocks, generate a first matrix including a plurality of values weighted by calculating the plurality. ade of weighted values based on the number of occurrences of the local default value and corresponding sizes of one or more macroblocks, assign a unique index to each of the plurality of weighted values, generate a second matrix of the single index by sorting the plurality of weighted values, and generate a third matrix including a plurality of classification distances, store a plurality of biometric templates of authorized personnel, compare the applicant's biometric template with the plurality of biometric templates of authorized personnel, generate a positive match signal in response to the identification of a matching the applicant's biometric template data to one of the plurality of authorized personnel biometric templates, and transmitting the positive match signal to a gateway to grant the applicant access to the point of entry.

BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Features believed to be characteristic of aspects of the disclosure are set forth in the appended claims. In the description that follows, like parts are marked throughout the specification and drawings with the same numbers, respectively. Drawing figures are not necessarily drawn to scale and certain figures may be shown exaggerated or generalized for clarity and brevity. The disclosure itself, however, as well as a preferred mode of use, other objects and advantages thereof, will be better understood by reference to the following detailed description of illustrative aspects of the disclosure when read in conjunction with the accompanying drawings.

[0012] FIG. 1 is an example of a real-time simultaneous identity verification and authentication system, in accordance with some aspects of the present disclosure.

[0013] FIG. 2 shows a perspective view of an example of a simultaneous real-time authentication and identity verification device, in accordance with some aspects of the present disclosure.

[0014] FIG. 3 shows a front view of an example of a simultaneous real-time authentication and identity verification device, in accordance with some aspects of the present disclosure.

[0015] FIG. 4 shows another perspective view of an example of a simultaneous real-time authentication and identity verification device, in accordance with some aspects of the present disclosure.

[0016] FIG. 5 is a block diagram of an example processing component of a simultaneous real-time identity verification and authentication device, in accordance with some aspects of the present disclosure.

[0017] FIG. 6 shows a flow diagram of a facial recognition method, in accordance with some aspects of the present disclosure.

[0018] FIG. 7(a) shows a facial image for a person, in accordance with some aspects of the present disclosure.

[0019] FIG. 7(b) shows a different facial image for the same person, in accordance with some aspects of the present disclosure.

[0020] FIG. 8 shows an example process for calculating the local binary pattern resource (LBP) in accordance with some aspects of the present disclosure.

[0021] FIG. 9 shows an example process for calculating the local ternary pattern resource (LTP), in accordance with some aspects of the present disclosure.

[0022] FIG. 10 shows positions of three example key features selected from one or more face images, in accordance with some aspects of the present disclosure.

[0023] FIG. 11 shows an example of a receiver operating characteristic (ROC) curve for testing a face database, in accordance with some aspects of the present disclosure.

[0024] FIG. 12 illustrates an example of asymmetric biometric encryption for confidentiality, in accordance with some aspects of the present disclosure.

[0025] FIG. 13 illustrates another example of biometric and asymmetric cryptography for authentication, in accordance with some aspects of the present disclosure.

[0026] FIG. 14 illustrates a schematic view of an example of an environment for implementing one or more doors for access control.

[0027] FIG. 15 illustrates an example of a computer system for implementing a data management method in accordance with aspects of the present disclosure.

[0028] FIG. 16 illustrates a block diagram of various exemplary system components, in accordance with aspects of the present disclosure.

[0029] FIG. 17 illustrates an example of a faceplate for identifying biometric templates, in accordance with aspects of the present disclosure.

[0030] FIG. 18 illustrates an example of the faceplate components of FIG. 17, in accordance with aspects of the present disclosure.

[0031] FIG. 19 illustrates another example of the faceplate components of FIG. 17, in accordance with aspects of the present disclosure.

[0032] FIG. 20 illustrates an example of a heat map, in accordance with aspects of the present disclosure.

[0033] FIG. 21 illustrates an example of LBP operation on measurement points of the heat map of FIG. 20, in accordance with aspects of the present disclosure.

[0034] FIG. 22 illustrates examples of sub-matrices.

[0035] FIG. 23 illustrates an example of a results table for sequence conversion, in accordance with aspects of the present disclosure.

[0036] FIG. 24 illustrates an example of a flowchart for converting a sequence, in accordance with aspects of the present disclosure.

[0037] FIG. 25 illustrates an example of a check calculation table in accordance with aspects of the present disclosure.

[0038] FIG. 26 illustrates an example of deep learning.

[0039] FIG. 27 illustrates the deep learning example in FIG. 27.

[0040] FIG. 28 illustrates a flowchart of a method for identifying biometric templates.

DETAILED DESCRIPTION

[0041] The following includes definitions of selected terms used herein. Definitions include various examples and/or forms of components that fall within the scope of a term and that can be used for implementation. The examples are not intended to be limiting.

[0042] A "processor", as used in this document, processes signals and performs general computation and arithmetic functions. The signals processed by the processor can include digital signals, data signals, computer instructions, processor instructions, messages, a bit, a bit stream, or other computation that can be received, transmitted and/or detected.

[0043] A "bus", as used in this document, refers to an interconnected architecture that is communicatively coupled to transfer data between computer components within a single or multiple systems. The bus can be a memory bus, a memory controller, a peripheral bus, an external bus, a crossbar switch and/or a local bus, among others. The bus can also be a vehicle bus that interconnects components within a vehicle through protocols such as CAN (Controller Area Network), LIN (Local Interconnect Network), among others.

[0044] A "memory", as used in this document, may include volatile memory and/or non-volatile memory. Non-volatile memory can include, for example, ROM (read-only memory), PROM (programmable read-only memory), EPROM (erasable PROM) and EEPROM (electrically erasable PROM). Volatile memory can include, for example, RAM (random access memory), synchronous RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM) and/or bus RAM of direct RAM (DRRAM).

[0045] As used in the specification and appended claims, the singular forms "a", "an" and "the" include plural referents, unless the context clearly indicates otherwise.

Ranges may be expressed herein as from "about", "substantially" or "approximately" a particular value and/or to "about", "substantially" or "approximately" another particular value. When such a range is expressed, another implementation includes from a particular value and/or to another particular value.

[0047] By "comprising" or "containing" or "including" it is meant that at least the named compound, element, particle or method step is present in the composition or article or method, but does not exclude the presence of other compounds, materials, particles, method steps, even if the other compounds, materials, particles, method steps have the same function as the one named.

[0048] It should also be understood that the mention of one or more method steps does not exclude the presence of additional method steps or intermediate method steps between those expressly identified steps. Likewise, it should also be understood that the mention of one or more components in a device or system does not exclude the presence of additional components or intervening components among those expressly identified components.

[0049] Biometric identification techniques generally refer to pattern recognition techniques that perform an applicant identification process by determining the authenticity of a specific physiological or behavioral characteristic possessed by the applicant. In some cases, biometric identification may be preferred over traditional methods involving passwords and personal identification numbers (PINs) for a variety of reasons. For example, with biometric identification, the person (eg applicant) to be identified normally must be physically present at the point of identification. Furthermore, identification based on biometric techniques avoids the need to remember a password or carry a token (ie, a security device used to gain access to a controlled access point of entry).

[0050] A type of texture-based local binary pattern ("LBP") resource describes facial information that produces desirable recognition results. The enhanced local ternary pattern ("LTP") feature can be an additional improvement over conventional LBP methods. The LBP and LTP features may not be sensitive to light and expression variations and are computationally efficient, but they also have shortcomings, such as information redundancy due to the correlation between the positive histogram and the negative histogram.

[0051] It is therefore desirable to contemplate simultaneous identity verification and real-time authentication techniques to create biometric signature data to provide keyless access to authorized requesters to a vehicle, building or the like with varying degrees of security, using various types biometrics data from authorized applicants. As discussed above, in some implementations of the present disclosure, biometric signature data may be interchangeable in a wide variety of applications. Therefore, in some examples of the present disclosure, the same biometric signature data for a person can be used to authenticate that person at one or more locations and for one or more applications. Furthermore, an example of a biometric system in the present disclosure allows biometric signature data to be altered based on a desired level of security. Thus, the type of biometric signature data that can be used for a specific application and/or related to a specific requester may vary depending on the level of security desired for that specific application and/or requester. Although some implementations discussed herein are discussed in the context of facial biometric data, those skilled in the art would understand that various implementations of the present disclosure can employ many types of biometric data, including, but not limited to, fingerprint data, scan data. of iris and retina, speech data, facial thermograms, hand geometry data, and the like.

[0052] In some implementations of the present disclosure, biometric data associated with the intended recipient (for example, a biometric template) may be obtained by means of a biometric sensor of a biometric-based access control system. As will be discussed below, variations in light, temperature, distance of the biometric sensor from a target can impact the quantity and quality of biometric data obtained through the biometric sensor. For example, variations in light intensity and angle can create shadows on an applicant's face,

making face recognition more difficult. If the biometric data to identify an applicant is obscured, more templates may be needed to properly authenticate the applicant, thus increasing the amount of biometric data needed. To reduce the undesirable impact of these environmental factors, the biometric sensor can utilize near infrared (IR) or ultraviolet (UV) light or a combination of IR and UV at the desired intensities. In one implementation, the method uses near infrared light. An infrared light emitting diode (LED) array can be used in the facial recognition device or biometric sensor to minimize the impact of surrounding lighting in capturing the facial uniqueness. The camera and LED array are packaged in a dedicated edge device (eg a faceplate) mounted in a location that requires verification and/or identification/analysis, such as a door that requires access control.

[0053] In some implementations, an access control system may use IR or new lighting and IR detection to identify facial features. IR or new IR illumination can penetrate the dermis of the face. The IR or new IR illumination can penetrate the dermis at 10 micrometers, 20 micrometers, 50 micrometers, 100 micrometers, 200 micrometers, 500 micrometers, 1 millimeter,

2mm, 5mm and/or 10mm. Other penetration depths are possible. Penetration depths may depend on body location, wavelength of infrared illumination and/or intensity of infrared illumination. Penetration can expose skin features that may be difficult to see in visible light, including (age spots, spider veins, hyperpigmentation, rosacea, acne, and porphyrins). The identification of these subdermal characteristics can be used to adjust/add to the applicant's unique identification. These features on the applicant's face may be unique because they are based on the applicant's exposure to nature and sun during the applicant's lifetime. Facial recognition based on subdermal features can identify the uniqueness of the face at the time of capture to provide opportunities for identification analysis. The number of subdermal features can increase over time with sun exposure and daily.

[0054] In another example, an access control system can use ultraviolet lighting and detection to identify facial features. Ultraviolet light can penetrate the dermis of the face. UV illumination can penetrate the dermis at 10 micrometers, 20 micrometers, 50 micrometers, 100 micrometers, 200 micrometers, 500 micrometers, 1mm, 2mm, 5mm and/or 10mm. Other penetration depths are possible. Penetration depths may depend on body location, wavelength of ultraviolet illumination and/or intensity of ultraviolet illumination. Penetration can expose skin features that may be difficult to see in visible light, including (age spots, spider veins, hyperpigmentation, rosacea, acne, and porphyrins). The identification of these subdermal characteristics can be used to adjust/add to the applicant's unique identification.

These features on the applicant's face may be unique because they are based on the applicant's exposure to nature and sun during the applicant's lifetime. Facial recognition based on subdermal features can identify the uniqueness of the face at the time of capture to provide opportunities for identification analysis. The number of subdermal features can increase over time with sun exposure and daily. The facial recognition system of the present disclosure can estimate a person's age based on the amount and nature of subdermal features. The access control system can also track the change in these features over time to confirm the individual's identity and establish lifestyle and daily routines based on interpretations of the subdermal features. Subdermal facial recognition can also increase the difficulty of creating a duplicate (eg, duplicate of a biometric template) of the face due to its elimination of reliance on facial features capable of being captured by standard visible wavelength photography and camera technology . The access control system can also further obfuscate the ultraviolet capture content by introducing time-sequenced cross polarization filters to the capture process that further eliminates the ability to present an artificial duplicate of the face to the access control system .

[0055] A benefit of the system in this disclosure includes allowing the replacement of a single credential system with PINs, passwords and multifactor authentication that is transparent to the requester. With this architecture in place, the system requester(s) can rely on a single credential management solution. The system of the present disclosure can support both logical and physical gates. In some implementations of the present disclosure, the system can provide protection at home and at work.

[0056] Aspects of the present disclosure may include a method referred to as "layered reinforcement". The method consists of taking the face image of the biometric sensor and superimposing several layers of pixel boxes of different sizes on the image. This layering of pixel boxes of different sizes has an amplifying impact on the analysis of the uniqueness of the face. The most exclusive areas of the face are enlarged. Areas that are most common between faces are not emphasized. As a result, layered reinforcement can improve algorithm performance while allowing the method to handle large numbers of users in multiple locations where the biometric sensor panel is deployed. The "layer reinforcement" of the method can allow processing the same number of requesters on an Advanced Reduced Instruction Set (ARM) Computer Machine processor located on the biometric sensor panel where the image is captured for the first time, thereby reducing the hardware and processing requirements and contributing to the accuracy and reliability of the method, as a network failure cannot prevent the biometric sensor panel from processing a face check.

[0057] Some aspects of this modality of the invention cover the use of a port (described below) to manage the data analyzed by the various algorithms to increase performance, decreasing false negative and false positive results through the following processes: hierarchical analysis of the box. pixel to create a binary tree of dominant features (that is, determine which is the most distinct feature); pixel box time domain analysis with heat maps (ie, determining features over time that are problematic due to overlap between individuals); and binary tree collision (signaling biometric signature data overlap for two individuals that can cause a false positive and proactively addressing).

[0058] Benefits to the system of the present disclosure include improved performance when accuracy requires reduction of false negative and false positive results. The enhancement also enables the benefits of 1:1 comparison in a 1:N environment as a potential replacement for video surveillance and comparison, thus opening the mass surveillance market to significantly increased accuracy.

[0059] With reference to FIG. 1, an example of an identification system 100 for simultaneous real-time identity verification and authentication for use in, for example, allowing access by an authorized applicant to a vehicle, building, or the like is illustrated in accordance with aspects of the present disclosure. .

[0060] It should be recognized that FIG. 1 is intended to describe aspects of disclosure to enable those skilled in the art. Other implementations may be used and changes may be made without departing from the scope of this disclosure.

[0061] The identification system 100 comprises a simultaneous real-time identity verification and authentication device 102, including at least one biometric sensor 104, a processor 106, memory 108, a screen 110 and input/output mechanism 112. identification system 100 can be used to secure or control access to a secured area, device, or information, such as an airport departure area, building, stadium, database, locked door, vehicle, or other controlled access assets/infrastructure .

[0062] The biometric sensor(s) 104 may include a camera, a fingerprint reader, retina scanner, facial recognition scanner, weight sensor, height sensor, temperature sensor body, motion sensor, heart rate sensor or any other sensor or device capable of detecting a biometric characteristic of a person. As shown in FIGS. 2-4, in an exemplary implementation of the present disclosure, the biometric sensor(s) 104 may be an optical sensor, such as a camera.

[0063] In some aspects, the biometric sensor(s) 104 may include an optical sensor that captures visual data. For example, the sensor(s)

biometric(s) 104 can be a camera that detects a requester's visual information, such as the person's facial features. A person's facial features can include the person's textures, complexion, bone structures, blemishes, birthmarks, contours, coloration of the person's face. The biometric sensor(s) 104 can capture the person's facial features and convert the visual information into detected digital information (as discussed below).

[0064] Processor 106 may be configured to compare information detected via biometric sensor(s) 104 with known characteristics of a person in an attempt to identify the person by means of biometric signature data. Processor 106 can include any number of processors, controllers, integrated circuits, programmable logic devices, or other computing devices. Processor 106 can be communicatively coupled with biometric sensor(s) 104 and other components of system 100 via wired or wireless connections to allow information to be exchanged between device 102 and external devices. 114 or systems (e.g., network 116) for enabling comparison of the stored biometric signature data with detected information obtained from the biometric sensor(s) 104.

[0065] Processor 106 can implement a computer program and/or code segments stored in memory 108 to perform some of the functions described herein.

The computer program may include an ordered list of executable instructions for implementing logic functions in device 102. The computer program may be incorporated into any computer-readable medium

(eg, memory 108) for use by or in connection with an instruction execution system, apparatus, or device, and executing the instructions.

Memory 108 may contain, store, communicate, propagate or transport the program for use by or in connection with the instruction execution system, apparatus or device.

Examples of memory 108 may include an electrical connection with one or more wires, a random access memory (RAM), a read-only memory (ROM), an erasable, programmable read-only memory (EPROM or Flash memory), a memory portable computer floppy disk or portable compact disk (CDROM) read-only option. The memory

108 may be integral with device 102, a self-contained memory, or a combination of both.

Memory 108 can include, for example, removable and non-removable memory elements such as RAM, ROM, Flash, magnetic, optical, USB, and/or other conventional memory elements.

[0066] In some aspects, the memory 108 can store the known characteristics of a number of people and various other data associated with the operation of the system 100, such as the computer program and code segments mentioned above, or other data to instruct the device 102 and the other device elements for carrying out the aspects described herein. The various data stored in memory 108 can be associated with one or more databases (not shown) to facilitate retrieval of the information, for example, via external devices 114 or network 116. Although memory 108 as shown in FIG. 1 is integrated in device 102, it should be recognized that memory 108 may be stand-alone memory located in the same cabinet as device 102, or it may be external memory accessible by device 102.

[0067] In one aspect, screen 110 can be configured to display various information related to system 100 and its underlying operations. For example, a notification device may be included (not shown) to indicate the detected biometric feature or the detected signal that fails to match the person's known characteristics and may include an audible alarm, a visual alarm and/or any other notification device. notification.

[0068] In one aspect, device 102 may also include input/output mechanism 112 to facilitate the exchange of data and other information between different components within device 102 or with various external devices 114 or systems over network 116.

[0069] For example, several I/O ports can be contemplated, including an SD (Secure Disk Digital) card slot, Mini SD card slot, Micro SD card slot or similar for receiving removable SD cards, Mini SD cards , Micro SD cards, or the like, and a USB port for attaching a communicatively attached USB cable to another computing device, such as a personal computer. In some aspects, the input/output mechanism 112 may include an input device (not shown) for receiving identifying information about a person to be identified. The input device may include a ticket reader, a credit card reader, an ID reader, a keyboard, a touch screen or any other device. In some other aspects, as described above, the input/output mechanism 112 can be configured to allow the device 102 to communicate with other electronic devices over the network 116, such as the Internet, a local network, a wide network, an ad. hoc or peer-to-peer network, or a direct connection like

USB, Firewire or BluetoothTM connection, etc. In one example, known characteristics about people can be stored and retrieved in remote databases or memory via the network 116. The input/output mechanism 112 can thus communicate with the network 116 using data transfer methods. wired or wireless data transfer methods such as WiFi (802.11), Wi-Max, BluetoothTM, ANT®, ultra wideband, infrared, cellular telephony, radio frequency, etc. In one aspect, the input/output mechanism 112 may include a cellular transceiver for transmitting and receiving communications over a communications network operable with GSM (Global System for Mobile Communications), CDMA (Multiple Access Code Division) or any others. known patterns.

[0070] Device 102 may also include a power supply (not shown) to provide electrical power to the various components contained therein. The power supply may include batteries, battery packs, power conduits, connectors and receptacles operable to receive batteries, battery connectors, or power cables.

[0071] In one aspect, device 102 can be installed and positioned at an access control entry point (not shown), such as a gate, locked door, etc. to prevent people from accessing certain areas until device 102 determines that the biometric characteristic and/or the signal matches the known characteristics. In some other aspects, as shown in FIGS. 2 to 4, device 102 can be a self-contained, compact, handheld, and portable device. In one example, such a self-contained, compact, handheld and portable device can be used to protect sensitive documents or information that are electronically stored and accessed on the Internet and/or an intranet. In some aspects, a simultaneous identity verification facility access unit can use biometric signature data to create interchangeable authentication for a variety of uses (eg office, home, smart phone, computer, premises).

[0072] With reference to FIG. 5, processor 106 in FIG. 1 can be configured to include, among other features, a detection module 502 and a recognition module 508 to provide real-time or near real-time identity verification and authentication with keyless access for authorized requesters to secure facilities or information. Detection module 502 may include a face detection module 504 for detecting facial features of an applicant. Detection module 502 may include an eye detection module 506 to identify the locations of an applicant's eyes. In some implementations, detection module 520 may include one or both of face detection module 504 and/or eye detection module 506. In some aspects, processor 106 may receive inputs (digital or analog) from ) sensor(s) 104.

[0073] FIG. 6 describes an example procedure of selecting key features from a database with a large amount of facial information and constructing a classifier that can distinguish different faces adequately. LBP and LTP can be used to provide a complete description of the face information and then, using an adaptive reinforcement learning algorithm ("adaboost"), you can select key features and build a classifier to distinguish different faces creating biometric signature data. This biometric signature data can be used to create universal verification and authentication that can be used for a variety of applications (eg computer, building access, smartphone, automobile, data encryption) with varying degrees of access and security (by eg network access, but with greater security for the requester's computer). In block 602, create a face sample database. For example, processor 106 and recognition module 508 can create a face sample database using unrecognized face samples. In one implementation, processor 106 and recognition module 508 can store, in memory 108, 1000, different people with each person showing 10 different postures and/or expressions.

[0074] In block 604, extract the LBP and LTP resources.

For example, detection module 502 and/or face detection module 504 can extract LBP and LTP resources from different blocks at different positions of each face sample.

[0075] In block 606, calculate the positive sample and the negative sample. For example, at least one of the face detection module 502, face detection module 504 and eye detection module 506 can calculate the absolute value distance of the feature for the same position of any two different images of a person and define this distance as a sample database of positive traits. In addition, the face detection module 502 and the face detection module 504 can, together or separately, calculate the absolute value distance of the feature for the same position of any two different images of different person and define that distance as sample database of negative characteristics.

[0076] In block 608, create an adaboost classifier. For example, face detection module 502 and face detection module 504 can select the most distinguishable key feature from the candidate feature database with adaboost and create a face classifier.

[0077] In block 610, generate the recognition result. For example, recognition module 508 can generate recognition result. Once there is a fixed dataset of macroblocks and the specific LPB ranging from 1 to 255 is determined, a value is assigned to that dataset unit based on the number of pixels within the block that satisfy that specific LBP.

For example, assuming a 10 x 10 macroblock in unit number 1 of 255 and LBP of 20, method 600 determines the number of pixels in the histogram that fall within that LBP of 20 in scalar value. Method 600 calculates the scalar value and then normalizes the value into a second array to solve the problem of determining the value within various macroblock sizes. The scalar value based on the known method was based on the macroblock size, where the maximum value can be from 100 to 1600, depending on the macroblock size. The scalar value in this second matrix can now be a percentage of the total pixels available in that macroblock to normalize the data for subsequent evaluation. Normalization makes the data not skewed based on macroblock size.

After normalization under the improved method, each unit of the dataset in this second matrix has the same weight.

This normalized data can be ranked to establish and assign a value from 1 to 2165, where the scale reflects the highest normalized value going to the top of the ranking. For example, if dataset 2000 had the highest value in the array, it would be assigned a value of 1 with descending value reflecting the datasets that have the lowest values. The second normalized matrix can then be converted to a simulated third DNA sequencing matrix, where the position is established within this third matrix based on its value in the previous ranking. The third matrix evaluates the position and calculates the differences between where the dataset appears in the sequence (for example, rank distance). This enhanced method analyzes features as opposed to scalar value based on uniqueness of features within the face and not just scalar values.

[0078] In block 620, test the face sample. For example, detection module 502 can optionally test face samples.

[0079] In block 622, extract the LBP and LTP resources.

For example, detection module 502 and/or face detection module 504 can extract LBP and LTP resources from different blocks at different positions of each face sample.

[0080] In addition, online recognition can include the following steps: (1) Calculate the main characteristic extracted from the offline stage of different blocks in different positions so that the face sample is identified.

(2) Calculate the key characteristic selected in step (1) with that of each face sample in the database and determine whether they belong to the same person or not. If the calculated distance is less than the defined threshold, it can be determined that they are the same person, otherwise it can be determined that they are not.

[0081] As shown in FIG. 8, an example process starts with creating a face database with different postures and different expressions. For example, you can include the images of, for example, 1000 different people and each person shows, for example, 10 images differently. FIG. 7(a) shows the different face image of the same person, and FIG. 7(b) refers to the different face image of a different person.

[0082] LBP and LTP can be used to describe the face. FIG. 8 shows an LBP resource calculation process, and FIG. 9 shows an LTP resource calculation process. In order to get so many features for describing face information, different block sizes can be divided into different face swatch positions.

For example, face size can be 100 x 100, block size can be w x h, w and h values can range from 2 to 100, and blocks from 7837 can be selected as a result. The identification system 100 can select LBP and LTP bin resources in different block sizes and make it the final candidate resource database.

[0083] The next step is to calculate positive samples and negative samples. The absolute value distance of the bin feature from the same position for different images of the same person can be calculated and defined as the positive sample. Also, the distance of the absolute value of the bin resource from the same position to different people can be calculated and defined as the negative sample. For example, the result may involve calculating 32356 positive samples and 58747698 negative samples.

[0084] After that, the key compartment resource that can distinguish all positive and negative samples among the large number of resource database can be selected with a learning algorithm. For example, one can choose the discrete adaboost learning algorithm to select the resource and build a classifier.

[0085] An example algorithm of using adaboost to sort may include the following computational steps:

1. Given f as the maximum negative sample error rate, d as the minimum positive sample correct rate, Ftar as the target negative sample error rate, and Dtar as the target positive sample correction rate that the classifier cascading must reach. P, N are the positive and negative database, respectively.

2. Set F0 = 1.0, D0 = 1.0 and i = 0;

3. When Fi > Ftar, i = i + 1, ni = 0, Fi = Fi-1; when Fi > fx Fi-1, ni = ni+1.

4. Calculate the strong classifier with n resources via adaboost in database P and N; calculate Fi and Di for current cascade classifier, adjust threshold value of current strong classifier until rate is not less than dxDi-1, N is defined as non-empty.

5. If Fi > Ftar, sort the currently obtained cascade classifier onto another negative sample image and determine, place the incorrectly determined image into N.

1) Given n computation sample (xi, yi),... , (xn, yn), yi = 0, 1, xi presents negative sample tag and positive sample tag, respectively.

2) Initialize weight ωI,i = 1/2m, 1/2l, where the number of positive samples is l and the number of negative samples is m.

3) Try t from 1 to T and perform the steps below repeatedly: a) Normalize the weighting b) Calculate a weak classifier hj for each characteristic fj and plot the error rate of that classifier c) Find the classifier ht with the lowest rate of ωt error between all weak classifiers calculated in the last step, d) Update the weighting between which

If x is classified correctly, ei = 0. Otherwise, ei = 1.

Get the strong classifier last: if then h(x) = 1, otherwise h(x) = 0. There,

[0086] FIG. 10 shows the position of the first three key features selected from the face image, taking online tests for face database of 100 people based on the selected offline features and the classifier.

[0087] FIG. 11 shows the recognition results for 100 people, where the X-axis represents the false acceptance rate, which means the incorrectly identified rate of face samples. The Y-axis represents the verification rate, which means the rate of correctly recognized face samples. As shown in FIG. 11, when the false acceptance rate is below 10-4, it can reach a 95% recognition rate. Face recognition in this example not only improves the robustness of the face sample, but also reduces its computational complexity, significantly improving face recognition.

[0088] Referring again to FIG. 5, in some aspects, detection module 502 may be configured to use, among other features, a face detection module 504 and an eye detection module 506 to process the acquired image of the person to be identified as follows.

Face Detection Module 504 - Inputs: Front face image acquired (gray image), face classifier - Outputs: Face frame positions and number of faces - Flow: a. Reduce acquired front face image to user defined size b. Calculate a full image of the reduced image c. Initialize a transverse window based on the size defined by the face classifier, for example, 20 x 20 d. Move the transverse window on the full image from left to right and then top to bottom with each movement distance corresponding to a user-defined distance. However, if the user-defined distance is zero, set the move distance to 1/20 the width of the transverse window.

and. Use the face classifier to determine if the current position of the transverse window defines a valid part of a face. In that case, save the current position of the cross-window rectangular frame as a result.

f. After traversing the entire entire image, increase the width and length of the traverse window by 1.1 times and repeat step e until the size of the traverse window exceeds the image size or the buffer allocated to save the results is used.

g. Return to Face Frame Position and Faces Eye Detection Module 506 - Inputs: Acquired Front Face Image (Gray Image), Face Frame Positions, Sorter for Left and Right Eyes, Left Eye Sorter, Eye Sorter right, left eye coarse detection classifier, right eye coarse detection classifier - Outputs: frame position for both eyes, left eye frame position and right eye frame position - Flow:

The. Obtain facial image from acquired frontal image b. If the user-defined classifier for left and right eyes is available, use the correspondingly defined face detection function to detect the left and right eyes of the acquired face image. Otherwise, estimate left and right eye positions based on experience.

ç. If user defined left/right eye course detection classifier for left/right eye is available, detect left/right eye in corresponding half of obtained facial image. Also, based on the gross detection result, determine whether the detected human individual is wearing glasses or not. If there are glasses, detect the facial image obtained and return with the results. If there are no glasses, continue to detect the obtained face image based on the coarse detection result and return the detection result without considering the presence of glasses.

(If the user-defined classifier for the individual wearing glasses is not available, detect the obtained face image without considering the presence of glasses).

d. If user-defined course detection classifiers are not available, determine if glasses are present by directly detecting the left/right half of the acquired facial image. If there are glasses, detect the facial image obtained and return with the results. If there are no glasses, continue to detect the obtained face image based on the coarse detection result and return the detection result without considering the presence of glasses. (If the user-defined classifier for the individual wearing glasses is not available, detect the obtained face image without considering the presence of glasses).

and. Turn back

[0089] In some aspects, processor 106 can further use, for example, a recognition module 508, to extract pertinent facial features obtained from detection module 502 to compare with known features and/or information from a number of people authorized as follows.

Recognition Module 508 Normalization - Inputs: image to be normalized (gray image), the coordinates of the left and right eye centers on the image axis (the origin is located in the upper left corner of the image). The meanings of the parameters: 1x refers to the x coordinate of the center point of the left eye (horizontal direction) in the output image divided by the width of the output image and 1y refers to the x coordinate of the center point of the left eye (vertical direction) in the output image divided by the height of the output image.

- Output: output image Feature extraction - Inputs: Normalized image (gray image) and resource types - Outputs: If output buffer is NULL, returns the dimensional degrees of the resource. Otherwise, assume the output buffer size equals the resource's dimensional degrees, write the image's resources into the buffer, and return the resource's dimensional degrees. Certain features are associated with a certain image size.

For example, feature #6 may require the image size of 100 by 100. Therefore, when the input image fails the corresponding defined image size requirement, a zero result may be returned.

Resource comparison - Inputs: two resources to be compared and the comparison method - Output: The smaller the comparison result (a floating point), the greater the similarity.

Getting Algorithm Information - Function: instruct the requestor to correctly assign algorithm parameters

- Input: Algorithm type based on usage context - Outputs: Algorithm parameter information including feature type, feature dimensional degrees, normalized image size, minimum distance, suggested range and distance type.

[0090] Many of the systems and methods described above can be used to create biometric signature data ("BSD") files that allow a system to identify and distinguish applicants with a high degree of accuracy. Various implementations of the present disclosure can employ BSD files to create an encryption/decryption key, thereby increasing the security of such keys. Examples of the present disclosure can generate asymmetric keys based on one or more BSD files such that, using a biometric sensor, a person's biometric measurement can act as that person's private key. Implementations of the present disclosure may also incorporate BSD files into digital rights management (DRM) security in such a way that the files cannot be decrypted or accessed by anyone other than the intended requestor or group of requesters, or encrypted by a so that the original owners, such as a company, can no longer access the files. Consequently, when using implementations of the present disclosure employing BSD files, when a file is accessed, the identity of the requester that accessed the file can be guaranteed.

[0091] BSD files can be generated by algorithmic analysis of data from an A/D IR and/or UV sensor. Therefore, many of these elements can be considered when constructing the asymmetric pair's private key (ie, analog and/or digital values). Thus, in some implementations of the present disclosure, various elements of a sensor can contribute real-time data or real-time analog data related to a recognition event in order to decrypt, thus ensuring a real-time event (ie, the actual measurement of the intended person) triggered authentication.

[0092] As shown in FIGS. 12-13, in accordance with some implementations of the present disclosure, messages may be sent as follows. A requester can register, for example, on a computer and create a public key for the requester. The requester then publishes the public key so that it is publicly known. Other people, systems, or entities can use the requestor's public key to encrypt messages to the requestor and send those messages to the requester. The requestor can decrypt the message using its private key created by one or more active BSD files associated with the requestor. Consequently, the sender of the message is assured that the requester is actually the person who decrypts the message, because the private key used to decrypt the message can be generated by the requestor's active biometric data. These encryption systems and methods provide substantial advantages over conventional systems and methods. For example, instead of simply combining anonymous asymmetric codes, using BSD filed in the encryption process, authenticate because it is inherent in the key itself.

[0093] Several implementations of this disclosure can also improve DRM. For example, DRM rules can allow additional content to be added to a file and additional rules to be required. DRM rules can be expressed in many rights management languages known in the art, including, but not limited to, XrML (Extensible Rights Markup Language), XMCL (Extensible Media Trading Language), ODRL (Extensible Rights Markup Language), open digital rights) and the like. Rules can specify the allowed actions (eg decrypt, encrypt, transfer, copy, edit, etc.). Rules can also specify the people authorized to perform actions and the conditions under which those actions are allowed. BSD files can be used to authenticate a requestor to determine if the requestor is one of the people specified in the rules.

[0094] Various systems and methods for biometric encryption and authentication may also find application in corporate settings, where, for example, employees use corporate devices for personal use as well as business, or when, for example, an employee uses a personal device and corporate digital assets are transferred to and from the personal device. By applying rules to documents that have certain digital signatures, there can be controllable segmentation between private and commercial interests. Both parties can have access to the parties to which they are entitled, but they may be prevented from accessing the parties that are not entitled to access. For example, possible apps include, but are not limited to, providing remote access, shopping, and conditional security.

[0095] In the case of remote access, various implementations of the present disclosure can generate BSD files used to authenticate a requestor, providing secure access to any remote network connection, ie VPN server, secure access to network email and/ or company-owned information from a remote device.

[0096] In addition, the biometric authentication techniques of the present invention can be used to make authenticated online purchases/transactions. For example, spending limits can be based on the applicant or an account's group profile. In order for an applicant to make a purchase, a system can use the biometric authentication techniques of this disclosure to authenticate that applicant's true identity to verify that the applicant is entitled to make the desired purchase.

[0097] Biometric authentication techniques can also be used to provide conditional security to various digital files. For example, files that contain sensitive information can only be accessed by authorized requesters, who can be authenticated using the requester's active BSD files.

Biometric encryption and authentication app for digital cinema

[0098] The biometric encryption and authentication techniques described here find many applications in the digital cinema industry. Movies are popular goods, especially before DVD release. To maximize production efficiency and distribution opportunities, films need to be accessed and managed by many different strata of applicants. People skilled in the art recognize that techniques capable of protecting digital assets in the digital cinema industry can be used to protect digital assets in almost any industry. Consequently, the principles described in this document are not limited to application in the digital cinema industry, but can be applied to any industry for a similar purpose.

[0099] Digital cinema security is seen as an end-to-end process, from production through distribution to consumption. SMPTE DC28, the body responsible for digital cinema standards, has identified five distinct areas of digital cinema: (1) capture; (2) production; (3) Master (cinema, house, video, trailers, test projections); (4) distribution (satellite, fiber, bundled); and (5) exposure (digital projector security). In each area identified by SMPTE DC28, a film is vulnerable to theft. To discourage theft, movies can be encrypted before distribution. The movies are then stored in their encrypted state at the theater until showtime. At show time, the movie is decrypted and unzipped. This decryption/decompression can take place on a server or on a projector.

[00100] In an exemplary SMPTE DC28 process, DC28.4 can represent the conditional access portions of the cinema delivery system. Encryption methods

Modem DRM proved to be sufficient to withstand unwarranted decryption attempts, but protecting the keys became an issue. From capture to exhibition and distribution, a movie is encrypted and decrypted multiple times. Accordingly, various biometric encryption and authentication techniques discussed in this document can be applied to one or more of the encryption, decryption, and authentication steps, in accordance with various implementations of the present disclosure.

[00101] With reference to FIG. 14, an example of a data management environment 1400 may rely on a first port 1402a, a second port 1402b, and a third port 1402c to route data over wired and/or wireless communication links. The first port 1402a can be implemented as a virtualized software-based port on a computer system. The second port 1402b can be a standalone device that routes data as described below. The third port 1402c can be a cloud-based port. Other architectures are possible.

[00102] In certain implementations, ports 1402 can perform various functions, including managing the movement of data to and from the biometric sensor, as described below, providing a networked solution that efficiently moves binary facial data between devices and when grouped (physical and virtual), providing a high availability solution for security projects. Port 1402 can receive accreditation data through an XML file structure. By actively monitoring and consuming XML files, port 1402 may be able to use a standardized universal interface agnostic to the programming language, operating system and data source connectivity type, support physical and logical access control requirements within the same method, provide an interface that supports simultaneous connectivity of different types of system and/or supports live or batch processing of credentials, includes immediate retrieval of a system of credentials through file replacement.

[00103] The 1400 environment can include one or more faceplates 1404, for example, faceplate-a 1404a, faceplate-b 1404b, faceplate-c 1404c, faceplate-d 1404d, faceplate-e 1404e, and faceplate -f 1404f. Faceplates 1404 can transmit data to ports 1402 to be routed to another device within environment 1400. For example, faceplate 1404a can send a person's biometric template and/or identity information to the third port 1402c . Third port 1402c can send one or more biometric templates to faceplate 1404a to perform matching operations.

[00104] Still referring to FIG. 14, in certain implementations, environment 1400 may include an access control server 1406a, an enterprise server 1406b, and a third-party server 1406c. Access control server 1406a may be communicatively coupled to one or more access controlled entry points 1408 via a wired or wireless communication link. Enterprise server 1406b may be communicatively coupled to storage device 1410. Storage device 1410 may be a network drive, local hard drive, flash drive, tape drive, or other suitable storage media.

[00105] Still referring to FIG. 14, during operations, for example, faceplate-b 1404b may receive a biometric template of a first applicant 1450a. The first requester 1450a may attempt to access one or more controlled access entry points 1408 controlled by access control server 1406a. The one or more controlled access entry points 1408 may include a vault, lock, secure door, secure gate, equipment or machinery, computing device, digital storage device, database or archive, for example. In some examples, one or more controlled access entry points 1408 may be a controlled access door or gate to an infrastructure, such as a warehouse, office building, restricted area, etc. The 1450a first applicant biometric template may include one or more of the 1450a first applicant's fingerprints, voice patterns, iris patterns, facial features, signature patterns, ear shapes, retinal patterns, movement, hand geometry.

[00106] In certain implementations, the b-face 1404b may extract the facial features of the first applicant 1450a and compare the facial features with the facial features of the authorized personnel. If the facial feature of the first applicant 1450a matches one of the facial features of the authorized personnel, the face-b 1404b can send a first positive match signal to the third port 1402c. The third port 1402c can route the first positive match signal to the access control server 1406a. Upon receiving the first positive match signal, the access control server 1406a may unlock one of the one or more access controlled entry points 1408 associated with the b-faceplate 1404b to allow access by the requestor 1450. first applicant 1450a does not match one of the facial features of the authorized personnel, the first applicant 1450a may not gain access to entry point 1408 associated with faceplate 1404b.

[00107] Still referring to FIG. 14, in certain examples, the d-faceplate 1404d may transmit a second positive match signal from a second requester 1450b at a first time and a third positive match signal at a second time from the second requester 1450b to the second port 1402b. The second port 1402b can route the second positive match and the third positive match to the first port 1402a, which can route the second and third positive match to the enterprise server 1406b. Enterprise server 1406b can use the second positive match and the third positive match to record access information associated with the second requester 1450b. For example, enterprise server 1406b might record, on storage device 1410, the first time as the arrival time of the second requester 1450b and the second time as the departure time on a business day. In another example, enterprise server 1406b may record the first and second times as the number of accesses to one or more controlled access entry points 1408 by the second requester 1450b. Enterprise server 1406b may also record, based on information from the second and third positive match signals, the facilities, equipment, files, locations, and information accessed by the second requester 1450b.

[00108] With further reference to FIG. 14, in certain implementations, faceplate 1404f may transmit a fourth positive match signal from the third requester 1450c to the first port 1402a. The first port 1402a can route the fourth positive match signal to the third-party server 1406c through a firewall 1412. Firewall 1412 can filter information transmitted through firewall 1412 and prevent malicious requesters from gaining unauthorized access. The fourth positive match signal may indicate to the third-party server 1406c that the third requester 1450c has gained access to one or more controlled access entry points 1408. For example, the fourth positive match may indicate to the third-party server 1406c that the third party requester 1450c accessed software that requires payment from the owner of the third party server 1406c.

[00109] Still referring to FIG. 14, in certain instances, a 1452 network administrator can install,

manage, update, maintain and/or control software on faceplates 1404, ports 1402, access control server 1406a, one or more controlled access entry points 1408, corporate server 1406b, storage device 1410 and/or firewall 1412 through a workstation 1414. Workstation 1414 may be a desktop computer, laptop, tablet, laptop, smartphone, or other suitable computer devices communicatively coupled via a wired or wireless connection wire to the third port 1402c. Network administrator 1452 can transmit software commands from workstation 1414 to third port 1402c to be routed to a destination. In some examples, the 1452 network administrator might update firmware on the 1404 faceplates. In other examples, the 1452 network administrator might install new software on the 1406a access control server.

In another example, the 1452 network administrator can perform maintenance operations, such as disk error checking and defragmentation, on the storage device.

1410. In another example, network administrator 1452 may lock or open one or more controlled access entry points 1408 in an emergency.

[00110] With reference to FIG. 14, in some implementations, an employee 1454, as a supervisor,

can use a requester terminal 1416 to access information through the second port 1402b. Requester terminal 1416 may be a desktop computer, laptop, tablet, laptop, smartphone, or other suitable computer devices communicatively coupled via a wired or wireless connection to the second port 1402b. In some examples, employee 1454 can download information, such as working hours, arrival time, access history, usage frequencies, using the requesting terminal 1416.

[00111] In certain examples, data exchanges within the 1400 environment may be encrypted. Data transmissions between ports 1402 and faceplates 1404 can use advanced TLS v1.2 communications with a proprietary key management framework.

Data transmitted over TLS v1.2 communications can be fully encrypted to remove the threat of exposing the data to unwanted parties. Data encryption can be further protected by protecting the generation of encrypted keys through the use of biometric data as the seed for key generation. As such, data exchanged within the environment can be difficult to access by unauthorized requesters.

[00112] In addition, ports 1402 can be used to manage data and blockchain creation for requesters. First port 1402a, second port 1402b, and third port 1402c can each include a blockchain wallet. The wallets will contain the applicant's personal credentials necessary for authentication on any device or application. Wallets can be tethered to ports 1402 to provide cyber security monitoring and to provide interaction between the wallet and facial recognition devices. The 1404 faceplates linked to the personal blockchain may be able to allow a transaction on the blockchain. Communications between blockchain wallets, a registry for blockchain transactions, and ports 1402 can use the blockchain protocol. The personal blockchain will also extend to devices that check more than one requestor, such as a bank ATM (not shown). Ports 1402 will use location tracking to move the link binary facial data and the link to the blockchain record between shared devices to improve transaction security and to manage the number of requesters maintained within each 1404 faceplate.

[00113] Aspects of the present disclosure may be implemented using hardware, software, a cloud network or a combination thereof, and may be implemented in one or more computer systems or other processing systems. In one aspect of the present disclosure, resources are directed to one or more computer systems capable of performing the functionality described herein. An example of such a computer system 1500 is shown in FIG. 15. One or more of ports 1402, servers 1406, firewall 1412, workstation 1414, and/or requesting terminal 1416 may be implemented based on computer system 1500.

[00114] Referring now to FIG. 15, computer system 1500 includes one or more processors, such as processor 1504. Processor 1504 is communicatively coupled to a communication infrastructure 1506 (e.g., a communication bus, crossbar, or network). Various aspects of software are described in terms of this example computer system. After reading this description, it will be evident to a person versed in the relevant technique(s) how to implement aspects of the disclosure using other computer systems and/or architectures.

[00115] Computer system 1500 may include a display interface 1502 that forwards graphics, text and other data from the communication infrastructure 1506 (or a frame buffer not shown) for display on a display unit 1530. computer 1500 also includes main memory 208, preferably random access memory (RAM), and may also include secondary memory 1510. Secondary memory 1510 may include, for example, a hard disk drive 1512 and/or a hard disk drive. removable storage 1514, representing a floppy disk drive, magnetic tape drive, optical disk drive, universal serial bus (USB) flash drive, etc. Removable storage unit 1514 reads and/or writes to removable storage unit 1518 in a well-known manner. Removable Storage Drive 1518 represents a floppy disk, magnetic tape, optical disk, flash drive, USB, etc., which is read and written to Removable Storage Drive 1514.

As will be recognized, removable storage unit 1518 includes a computer-usable storage medium having computer software and/or data stored thereon.

[00116] Alternative aspects of the present disclosure may include secondary memory 1510 and may include other similar devices to allow computer programs or other instructions to be loaded into computer system 1500. Such devices may include, for example, a removable storage unit 1522 and a 1520 interface. Examples of such may include a program cartridge and a cartridge interface (such as found in video game devices), a removable memory chip (such as erasable programmable read-only memory (EPROM) or memory-only memory. programmable read (PROM)) and associated socket, and other removable storage units 1522 and interfaces 1520, which allow software and data to be transferred from removable storage unit 1522 to computer system 1500.

[00117] Computer system 1500 may also include a communication interface 1524. Communication interface 1524 allows software and data to be transferred between computer system 1500 and external devices. Examples of 1524 communication interface may include a modem, a network interface (such as an Ethernet card), a communication port, an International Association of Personal Computer Memory Card (PCMCIA) card slot and card, etc. The software and data transferred via the 1524 communication interface are in the form of signals 1528, which may be electronic, electromagnetic, optical, or other signals capable of being received by the 1524 communications interface. 1524 communications through a communication path (eg channel)

1526. This path 1526 carries signals 1528 and can be implemented using one or more of a wire or cable, fiber optic, telephone line, cellular link, RF link, and/or other communication channels. In this document, the terms "computer program medium" and "computer usable medium" are used to refer generally to media such as a 1518 removable storage drive, a hard disk installed in the 1512 hard disk drive, and signs

1528. These computer program products provide software for computer system 1500. Aspects of the present disclosure are directed to such computer program products.

[00118] Computer programs (also referred to as computer control logic) are stored in main memory 1508 and/or secondary memory 1510. Computer programs can also be received through communication interface 1524. Such computer programs, when performed, enable computer system 1500 to perform the features in accordance with aspects of the present disclosure, as discussed in this document. In particular, computer programs, when executed, allow processor 1504 to execute resources in accordance with aspects of the present disclosure.

Therefore, such computer programs represent controllers of computer system 1500.

[00119] In one aspect of the present disclosure where the method is implemented using software, the software can be stored in a computer program product and loaded into computer system 1500 using removable storage drive 1514, hard drive 1512, or computer interface. 1520 communication. The control logic (software), when executed by the 1504 processor, causes the 1504 processor to perform the functions described herein.

In another aspect of the present disclosure, the system is implemented primarily in hardware using, for example, hardware components such as application specific integrated circuits (ASICs). Implementing the hardware state machine in order to perform the functions described in this document will be evident to those skilled in the relevant technique(s).

[00120] FIG. 16 illustrates a block diagram of various example system components, in accordance with an aspect of the present disclosure. Fig. 16 shows a communication system 1600 usable in accordance with aspects of the present disclosure. Communication system 1600 includes one or more accessors 1660, 1662 and one or more terminals 1642, 1666. In one aspect, data for use in accordance with aspects of the present disclosure is, for example, input and/or accessed by accessors 1660 , 1662 via terminals 1642, 1666, such as personal computers (PCs), minicomputers, mainframe computers, personal computers, telephone devices, or wireless devices such as personal digital assistants ("PDAs") or portable wireless devices attached to a 1643 server, such as a PC, minicomputer, mainframe computer, microcomputer or other device with a processor and a data repository and/or connection to a data repository, through, for example, a 1644 network such as the Internet or an intranet and 1645 couplings , 1646, 1664.

Couplings 1645, 1646, 1664 include, for example, wired, wireless and/or fiber optic links. In another example variation, the method and system according to aspects of the present disclosure operate in a standalone environment, such as a single terminal.

[00121] Turning now to FIG. 17, an example of a faceplate 1404 that is configured to perform access control can analyze a biometric template of an applicant 1450 to determine if the applicant 1450 is authorized to gain access to an entry point (not shown) associated with the card. faceplate 1404. The faceplate 1404 may include an optical sensor 1702, a light source 1704, a display 1706, a keyboard 1708, and a scanner 1710.

[00122] In some implementations, the 1702 optical sensor can be configured to capture still or moving images. For example, the optical sensor 1702 can capture fingerprints, iris patterns, facial features, signature patterns, ear shapes, retinal patterns, motion, and/or hand geometry of the applicant 1450. In some examples, the optical sensor 1702 may be a wideband camera configured to detect electromagnetic radiation with wavelengths ranging from 200 nanometers (eg, soft UV) to 2000 nanometers (eg, near infrared (NIR)). In a particular example, optical sensor 1702 is configured to detect radiation between 700 to 900 nanometers. In certain implementations, optical sensor 1702 may include a motion sensor configured to detect people approaching faceplate 1404.

[00123] In certain examples, the 1702 optical sensor may include a wide-angle lens (eg, like a fisheye lens) used to provide vertical area coverage to capture faces across the full range of human heights (as well as address the Disabilities Act Requirements) and to provide horizontal coverage of the entire area of an access point to handle more than one person accessing the secure area in an authentication. In other examples, optical sensor 1702 may employ a high resolution charge-coupled device (CCD) array (e.g., megapixel per square inch). The high resolution matrix can provide the ability to identify faces at a greater distance from the sensor. The 1404 faceplate can take advantage of the increased identification distance to pre-identify applicants in queuing situations. Potential applicants can be identified as they enter a queue area and are placed on a priority list to speed up acknowledgment at the controlled access point of entry and unloading from the queue. Pre-identifying and prioritizing individuals in the point-of-access identification process allows for a high volume of transfer at the point of entry, reducing identification time during the identification and authentication process.

[00124] In certain implementations, the 1704 illumination source can emit electromagnetic radiation with wavelengths ranging from 200 nanometers to 2000 nanometers. In certain examples, the illumination source 1704 can emit non-visible radiation between 700 to 900 nanometers and/or 200 to 300 nanometers. Light source 1704 can emit radiation to illuminate applicant 1450's bodily characteristics and patterns used for biometric analysis (analyzing biometric model to determine access rights). The emitted radiation may collide with a portion of a requester's body 1450 and reflect off the portion of the body. The reflected radiation can be captured by the 1702 optical sensor for biometric analysis.

[00125] In some implementations, display 1706 may present useful information to requester 1450. For example, display 1706 may show one or more images of a face 1730 of requester 1450, captured by optical sensor 1702, to assist requester 1450 in the 1730 face alignment during the biometric analysis. In another example, screen 1706 may notify requester 1450 of an entry point status associated with faceplate 1404 (eg, locked, temporarily unavailable, normal operations, under maintenance). In yet another example, screen 1706 may display information such as time, date, weather, current location, etc.

[00126] Still referring to FIG. 17, keypad 1708 may allow applicant 1450 to enter numbers, symbols, and alphabets on faceplate 1404. In one example, applicant 1450 may enter a password in addition to biometric analysis to gain access to the entry point.

[00127] In some implementations, the 1710 scanner may be a radio frequency identification (RFID) scanner, a proximity card scanner (eg HIDTM card scanner), a contact card scanner, or a magnetic card scanner . In one example, scanner 1710 may send an interrogation signal to a proximity card (not shown) having a coil and an integrated circuit with a programmable or non-programmable identification sequence. The interrogation signal can be “absorbed” by the coil and can power the integrated circuit. In response, the powered integrated circuit sends a response signal including the identification string back to scanner 1710 via the coil. The 1710 scanner, in turn, analyzes the identification string to determine whether or not to grant access. The identifying sequence can be one or more numbers, alphabets, symbols and/or a combination thereof.

[00128] Still referring to FIG. 17, in one implementation, requester 1450 may approach faceplate 1404 during operations. The optical sensor 1702 can detect the requester 1450. In response to detection, the light source 1704 can emit incident NIR radiation 1760 towards the requester 1450. The incident NIR radiation 1760 may collide with the face 1730 of the requester 1450 and reflect off the face 1730 of the applicant

1450. In some implementations, optical sensor 1702 may detect NIR radiation detected 1762 from the surface of face 1730. NIR radiation detected 1762 may include incident NIR radiation reflected 1760 and/or NIR radiation emitted from the requestor 1450 due to thermal heating (or ie, blackbody radiation). The intensity and distribution of the detected NIR radiation 1762 may depend on the intensity and angle of the incident NIR radiation 1760, the contour of the face 1730, the angle of detection by the optical sensor 1702, and other factors. Faceplate 1404 can use the NIR radiation detected 1762 to build a facial model (the "NIR heat map") of the requester 1450. Faceplate 1404 can compare the constructed NIR heat map with existing models stored on it (details described below). If faceplate 1404 detects a match, faceplate 1404 can allow requester 1450 to access the entry point (as described above).

[00129] In some instances, an NIR heatmap generated by detecting NIR radiation can be resistant to changes in ambient lighting. As ambient lighting fluctuates (eg changes in luminance, color, color temperature, angle of illumination), an NIR heatmap constructed using NIR radiation detection can remain constant enough to avoid false acceptance or false rejection. For example, the applicant's 1450 NIR heatmap constructed via detection of NIR radiation under a "bright" condition (eg 1000 lux) may be substantially identical to the NIR heatmap constructed via detection of NIR radiation under a condition "dark" (eg 100 lux). In another example, the applicant's 1450 NIR heat map constructed via detection of NIR radiation under substantially blue light (eg, CIE coordinates x = 0.153, y = 0.100) may be substantially identical to the NIR heat map constructed via radiation detection NIR under substantially white light (eg CIE coordinates x = 0.30, y = 0.33).

[00130] Still referring to FIGs. 14 and 17, in other implementations, an NIR heat map generated by detecting NIR radiation can improve owner privacy. For example, storage device 1410 can store employee NIR heat maps and associated sensitive information (eg, dates of birth, email account passwords). If an unauthorized person gains access to the NIR heat maps and associated confidential information, the unauthorized person may not be able to exploit the stolen confidential information because it can be difficult to identify employees based on the NIR heat maps. Since NIR heatmaps are constructed using, for example, detected NIR radiation, they may be unrecognizable because a person's NIR heatmap can be drastically different from the visual image of the same person's face.

[00131] In another implementation, the requestor 1450 may approach the faceplate 1404 during operations. Optical sensor 1702 can detect the requestor

1450. In response to detection, illumination source 1704 may emit incident UV radiation 1760 (eg, electromagnetic radiation with wavelengths between 315 to 390 nanometers) toward requester 1450. Incident UV radiation 1760 may impinge on the face 1730 of the applicant 1450 and reflect on the face 1730 of the applicant

1450. In certain instances, incident UV radiation 1760 may penetrate the surface of the face 1730 and reflect the subdermal features of the face 1730 (eg, dermis, subcutaneous tissue, muscle, imperfections). In some implementations, optical sensor 1702 can detect UV radiation detected 1762 from the surface and/or subdermal features of face 1730. UV radiation detected 1762 may include reflected incident UV radiation

1760. The intensity and distribution of the UV radiation detected 1762 may depend on the intensity and angle of the incident UV radiation 1760, the contour of the face 1730, the angle of detection by the optical sensor 1702, and other factors. The faceplate 1404 can use the UV radiation detected 1762 to build a facial model (the "UV heat map") of the requester 1450. The faceplate 1404 can compare the built UV heat map with the existing models stored in it (details described below). If the faceplate 1404 detects a match, the faceplate

1404 can allow requestor 1450 to access the entry point (as described above).

[00132] In some examples, a UV heat map generated by detecting UV radiation can be resistant to changes in ambient lighting. As ambient lighting fluctuates (eg changes in luminance, color, color temperature, lighting angle), a UV heat map constructed using UV radiation detection can remain constant enough to avoid false acceptance or false rejection. For example, the UV heat map of applicant 1450 constructed by detecting UV radiation under a "bright" condition (eg 1000 lux) may be substantially identical to the UV heat map constructed by detecting UV radiation under a "dark" condition (eg 100 lux). In another example, the UV heat map of the applicant 1450 constructed via UV radiation detection under substantially blue light (eg CIE coordinates x = 0.153, y = 0.100) may be substantially identical to the UV heat map constructed via radiation detection NIR under substantially white light (eg CIE coordinates x = 0.30, y = 0.33).

[00133] Still referring to FIGs. 14 and 17, in other implementations, a UV heat map generated by detecting UV radiation can improve owner privacy. For example, storage device 1410 can store employee UV heatmaps and associated sensitive information (eg, dates of birth, email account passwords). If an unauthorized person gains access to the UV heat maps and associated confidential information, the unauthorized person may not be able to exploit the stolen confidential information because it can be difficult to identify employees based on the UV heat maps. Since UV heat maps are constructed using, for example, detected UV radiation, they may be unrecognizable because a person's UV heat map can be drastically different from the visual image of the same person's face. A heatmap can be a rendering of the dataset that visualizes the consistency of position within the three arrays. If the heat map is compromised, it may be more difficult to obtain biometric information used to identify the individual.

[00134] In some examples, the faceplate 1404 can generate a visible light heat map based on the detected visible light reflected from the face 1730 of the requestor

1450.

[00135] In certain implementations, the 1450 requestor may be required to provide a valid password, personal identification number (PIN) and/or HIDTM card to be used in conjunction with the built-in heatmap to gain access to the point of input associated with the faceplate 1404.

[00136] In other implementations, the faceplate 1404 may include a microphone (not shown in FIG. 17) to perform speech recognition.

[00137] In some examples, display 1706 may show face 1730 of requester 1450 as represented by optical sensor 1702. Display 1706 may include alignment marks (not shown) to help requester 1450 align face 1730 with respect to 1702 optical sensor. This alignment process can minimize false acceptances and false rejections due to misalignment. In certain implementations, faceplate 1404 may be a pocket mobile device powered by rechargeable batteries.

[00138] With reference to FIG. 18, in some implementations, faceplate 1404 may include a processor 1802 having a communication module 1852 configured to communicate with ports 1402 and other faceplates 1404, as described in this disclosure.

Communication module 1852 may be implemented as hardware in processor 1802, for example, as software code executed by processor 1802 or a combination thereof. The 1802 processor may also include an 1854 security module configured to encrypt and/or decrypt data.

Security module 1854 can be implemented as hardware on the 1802 processor, for example, as software code executed by the processor.

1802 or a combination thereof.

Processor 1802 further includes an algorithm module 1856 for building and comparing biometric models as described throughout this disclosure.

Alternatively, communication with the port can be facilitated by a processor on the control panel.

Algorithm module 1856 may be implemented as hardware in processor 1802, for example, as software code executed by processor 1802 or a combination thereof.

Processor 1802 may further include a parallel computing module 1858 for performing distributed processing.

The parallel computing module

1858 can be implemented as hardware on the processor

1802, for example, as software code executed by the 1802 processor or a combination thereof.

Processor 1802 may include one or more processors or cores and may be implemented as a semiconductor processor, graphics processing unit, an array of field-programmable gates, a programmable logic device, a processing cluster, an application-specific integrated circuit, or other suitable architectures.

[00139] The faceplate 1404 includes a memory

1804. Memory can be static or dynamic memory, such as flash memory, random access memory, magnetic memory, or semiconductor memory. Memory 1804 can include external memory, such as cloud storage. Memory 1804 may include or store applications and/or computer executable code. The faceplate 1404 further includes a modem 1808 to communicate with ports 1402 and other faceplates 1404 and may operate in cooperation with the communication module 1852. The faceplate 1404 also includes a RAM 1806, such as static or random access memory. dynamic (RAM). Faceplate 1404 may also include an input/output (I/O) device 1810 communicatively coupled to display 1706, optical sensor 1702, keyboard 1708, and scanner 1710. Components within faceplate 1404 may be interconnected by a 1812a inner bus panel. Processor 1802, memory 1804, RAM 1806, and internal bus 1812a can be arranged on an 1820 processor board.

[00140] Referring now to FIG. 19, another example of faceplate 1404 may include a series of processing boards 1820 interconnected by an external bus 1812b. Each processor board 1820 may include processor 1802, memory 1804, RAM 1806, and internal bus 1812a. The data distributed between the processing boards 1820 can be distributed by a controller 1814 via the external bus 1812b. In a non-limiting example, faceplate 1404 can download (from ports 1402) 240,000 biometric templates of potential applicants. The controller 1814 can distribute the 240,000 biometric templates evenly or unevenly among the processing boards 1820 so that each processing board 1820 can store, in its memory 1804, 30,000 biometric templates. In some implementations, each processor board 1820 may store the same or a different number of biometric templates.

[00141] Still referring to FIG. 19, in an example, during operations, faceplate 1404 may construct a heatmap (eg, UV or NIR) of the requestor 1450 based on the detected NIR radiation 1762. Controller 1814 may distribute copies of the constructed heatmap for each of the 1820 processing boards.

In some implementations, the 1820 processor boards can simultaneously compare the duplicated heat maps with locally stored biometric templates (eg, 30,000 stored on each 1820 processor board). Although the current example of the faceplate 1404 shown in FIG. 19 include eight processing boards

1820, other processing board numbers 1820 can also be used. For example, faceplate 1404 may include 2, 4, 6, 8, 12, 16, 32, or 64 processing boards 1820.

[00142] In one implementation, the faceplate 1404 may rely on remote processing cards (not shown) to perform the distributed computing described above. For example, the faceplate 1404 can send the duplicated heatmaps to the remote processing boards to jointly and simultaneously implement the algorithm (described below) for combining the duplicated heatmaps (or the numerical representation of the duplicated heatmaps) to known biometric models. Processing cards can be inside other faceplates of the network. In certain examples, the faceplate 1404 can send the duplicated heatmaps to a Beowulf cluster. The cluster design can employ both interprocess and interprocessor protocols to share processing tasks for the same application between the two processors and the processing cores within those processors. In the case of facial recognition, processor intensive operations such as face checking can use multiple processors or multiple cores to complete the operation within an acceptable period of time. In some cases, large multi-core processors can be used for this type of operation. In other cases, operation can be spread across multiple advanced RISC machine processors (“ARM”) to achieve the same performance as the single multi-core processor without the high hardware cost and potential for a single point of failure.

[00143] Referring back to FIG. 14, in a non-limiting example of distributed parallel comparison, the first port 1402a can download 600,000 biometric templates of potential requesters (eg, employees and contractors) from the 1406b enterprise server.

First port 1402a can distribute 100,000 biometric templates each of faceplate-e 1404e and faceplate-f 1404f, and 200,000 biometric templates each of second port 1402b and third port 1402c. Then the second port 1402b can distribute 100,000 biometric templates for each of faceplate-c 1404c and faceplate-d 1404d and the third port 1402c can distribute

100,000 biometric templates for each of the faceplate-a 1404a and faceplate-b 1404b. Consequently, the 600,000 biometric templates downloaded from the 1406b enterprise server can be evenly distributed across the 1404 front panels (that is, 100,000 non-overlapping templates each). In some implementations, when b-faceplate 1404b builds a heatmap (UV or NIR) of the first requester 1450a, b-faceplate 1404b can duplicate the built-in heatmap and distribute the duplicated heatmaps to the faceplate -a 1404a, face-c 1404c, faceplate-d 1404d, faceplate-e 1404e, and faceplate-f 1404f (via one or more ports 1402). The 1404 faceplates can compare the duplicated heat map of the 1450a first requester with the

100,000 templates stored locally to determine a match. Faceplate-a 1404a, faceplate-c 1404c, faceplate-d 1404d, faceplate-e 1404e, and faceplate-f 1404f can send (via one or more ports 1402) the result of the comparison back to faceplate-b 1404b. The faceplate-b 1404b can gather the results and determine whether to grant access to the first requester 1450a. In other implementations, biometric templates may be unevenly distributed among the faceplates 1404.

[00144] Referring now to FIG. 20, faceplate 1404 can generate a heat map 2000 based on measuring the detected radiation intensity 1762 of face 1730 of requester 1450. Heatmap 2000 can include a base 2010 matrix of one or more measurement points

2002. The 2002 measurement points may include a value indicative of the intensity of radiation detected in 1762 at a particular measurement point location. For example, the 2002a measurement point-a may indicate an intensity scale value of 2 corresponding to a 2004a region (eg, black hair) around the 1730 face of the requestor.

1450. Measuring point-b 2002b can indicate an intensity scale value of 50 that corresponds to a background color. The 2002c-c measurement point may indicate an intensity scale value of 21 that corresponds to a 2004c region (eg, cheeks) on the 1730 face of the requestor.

1450. In certain examples, the intensity scale can range from 0 (no reflection) to 100 (maximum reflection capable of being detected by the front plate optical sensor 1702

1404. In one implementation, the intensity scale can measure an absolute intensity (eg, brightness) of the 2002 measurement points. waveform (eg UV, NIR, red, green, blue).

[00145] Although heat map 2000 in FIG. 20 show the 2010 base matrix of 10 x 12 measurement points 2002 through the face 1730 of the requester 1450, another measurement point density may be possible for the base matrix 2010. For example, the faceplate 1404 can generate another map measurement using 100 x 100 points measurement across face 1730 of requester 1450. In another example, faceplate 1404 can generate a heat map using 500 x 500 measurement points. Other measuring point densities are also possible and may depend on desired accuracy, computational constraints, amount of data storage, etc.

[00146] In some implementations, the faceplate 1404 may remove 2002 measurement points indicating the background.

[00147] Turning now to FIG. 21, in one implementation, faceplate 1404 (or processor 1802) can apply LBP operation to one or more of the 2002 measurement points within the 2010 base matrix. For example, faceplate 1404 can perform LBP at the point of d-measurement 2002d, which includes an intensity value of 41. The LBP column for the 2002d-measurement point 2002d is 01110100, based on the range of 1 (that is, running LBP using the immediate neighbors, with a distance of 1 cell , from the measurement point- d 2002). Faceplate 1404 can track the number of LBP columns for the remaining 2002r measurement points, as shown in a 2100 table. In some examples, faceplate 1404 can calculate 24 of the 2002 measurement points as having LBP column of 00000000, 3 having the LBP sequence of 00000001, 11 having the LBP sequence of 00000010 and 0 having the LBP sequence of 00000011, etc., as indicated in table 2100. Table 2100 can include 256 entries for the possible sequences (ie 8 bits ). In some examples, table 2100 may include 255 entries,

eliminating the entry for the "all white" column (11111111) or the "all black" column (00000000). In some implementations, the data in table 2100 can be plotted as a histogram indicating the number of occurrences (eg 2002 measurement points) for possible LBP sequences.

[00148] In some implementations, based on the distribution of LBP columns in table 2100, faceplate 1404 may determine one or more unique features. The one or more unique resources may be non-zero LBP sequences occurring less than a threshold frequency (eg 5), such as LBP sequences 00000001, 01110100 and 11111100. In other implementations, one or more unique resources may be LBP sequences nonzero occurring less, such as LBP 01110100 sequences. In some examples, the faceplate 1404 can track the location of the 2002 measurement points having one or more unique features. For example, faceplate 1404 can track the coordinate of LBP columns. In some implementations, one or more unique features may be non-zero LBP sequences occurring more.

[00149] Turning now to FIG. 22, in certain implementations, faceplate 1404 may divide base array 2010 into one or more sub-arrays. Each sub-matrix can include a macroblock of measurement points. For example, the faceplate 1404 can divide the 2010 base matrix into a 2202 5x5 submatrix, a 2204 3x3 submatrix, and a 2206 6x6 submatrix. When calculating the LBP sequences for the 2002 measurement points within the 5 submatrix x 5 2202, the faceplate 1402 can calculate the LBP sequences of the 2002 measurement points within the 5 x 5 2202 sub-array using a range of 3 (i.e., calculate LBP sequences using neighbors 3 cells away). Likewise, when calculating the LBP sequences for the 2002 measurement points within the 3 x 3 2204 sub-array, the faceplate 1402 can calculate the LBP sequences of the 2002 measurement points within the 3 x 3 2204 sub-array using an interval of 2. By calculating the LBP sequences for the 2002 measurement points within the 2206 6x6 sub-array, the 1402 faceplate can calculate the LBP sequences of the 2002 measurement points within the 2206 6x6 sub-matrix using an interval of 4.

Other sizes for the base array, sub-arrays and extensions are possible as determined by faceplate 1402.

[00150] In one example, a 100 x 100 base matrix can be divided into sub-matrices of six different sizes: 10 x 10, 15 x 15, 20 x 20, 30 x 30, 35 x 35 and 40 x 40. Inside each In one of these sub-matrices, the LBP sequence can be calculated with a different interval around the measurement point being calculated to characterize the texture/slope of the area around the cell in different coverage areas. The range in pixels of the 2002 measurement point to be calculated for each of the neighboring cells for each of the sub-matrix sizes can be 3, 9, 15, 21, 27, and 33. In another example, a 200 x 200 base matrix can include 6 different sub-matrices with sizes of 10x10, 20x20, 35x35, 50x50, 65x65, 80x80 and 100x100. The pixel amplitude of the 2002 measurement point to be calculated for each one of the neighboring cells for each of the sub-array sizes can be 3, 5, 7, 10, 25, and 40. The sub-arrays may overlap in some cases. In certain implementations, the faceplate 1404 can apply LTP calculations on the 2002 measurement points.

[00151] Referring now to FIG. 23, in certain examples, the faceplate 1404 can convert the list of binary resources derived from the heat map into a data string that is extensible and capable of being tailored to the unique characteristics of each requester 1450, while still providing a comparison methodology of objects and correspondence/verification. A binary resource can contain three characteristics: the size of the subarray/macroblock, the location of the subarray/macroblock in the heatmap, and the Local Uniform Binary Pattern (ULBP) assigned to the binary resource. A macroblock can be a "subpicture" of the image of face 1730 of requester 1450. For example, a macroblock of size 10 is a 10x10 subpicture (ie, 100 measurement points 2002) of the image of face 1730. In one Non-limiting example, the location of the macroblock in the image can be defined by the position of the first pixel in the macroblock located in the upper left corner of the macroblock. The position of this pixel is defined by two values, its distance from the left edge of the face image (x dimension) and its distance from the upper edge of the face image (y dimension). The final characteristic is ULBP, which is a mathematical methodology for establishing a scalar value for the edge and texture characteristics of a pixel starting with the upper-left corner of the face.

1730. For each pixel within the macroblock, an ULBP calculation is performed. If the calculated ULBP matches the ULBP defined for that binary resource, the value of the binary resource is incremented by one. Therefore, the maximum value of a binary resource is the defined macroblock size (10 x 10 is 100) and the minimum value is zero. A macroblock can be used in more than one binary resource with different ULBP definitions. The location of a macroblock can also be used in more than one binary resource with a different macroblock size and ULBP definition.

[00152] Still referring to FIG. 23, a table 2300 can illustrate an example of the results of the sequence conversion process for macroblocks with sizes of 10x10, 35x35, 20x20, 30x30, 15x15, 40x40 and 15x

15. Three matrices can be used to generate a biometric template. The first matrix can be a scalar value in the normalized matrix. The second array can be sorted so that binary resource 2000, which in this example has the highest value, goes to the top of the second array at position 1. The third array compares the difference in position of binary resource 2000 in sorting instead of scalar value. If the position within the array for binary resource 2000 remains the same, the results will reflect a value of 0 representing the smallest possible change in position and therefore the highest significance value in the authentication method.

[00153] Turning now to FIG. 24, a method 2400 of converting a string can identify the most unique features used for identification.

[00154] In block 2402, obtain the heat map of the face. For example, processor 1802 and/or communication module 1825 can obtain the heat map of face 1730.

[00155] In block 2404, assign a unique index to each macroblock/ULBP combination. For example, algorithm module 1856 can assign a unique index (eg 4) to each macroblock/ULBP combination (eg 30 x 30 macroblock) and analysis based on uniqueness of features within the face rather than values scalars.

[00156] In block 2406, build a first array of scalar values for each macroblock/ULBP combination in the master schema referenced by the schema. For example, algorithm module 1856 may construct a first matrix of scalar values for each macroblock/ULBP combination in the master schema referenced by the schema. The first matrix includes the number of pixels that fall into that ULBP within that macroblock. By analyzing the LBP of a pixel, the faceplate 1404 can perform the normal LBP calculation and get a histogram that is the value from 0 to 255, and if the result is the highest value (standard LBP formula that takes pixel and takes 8 pixels around it and each pixel has a different binary value), the result is given a value of 1. If some other LBP is larger than that pixel, it will have a value of 0.

[00157] In block 2408, weight each scalar value by the macroblock size. For example, algorithm module 1856 may weight each scalar value (eg, 380) by macroblock size (eg, 30x30).

[00158] In block 2410, sort the first array of scalar values in descending order. For example, algorithm module 1856 can sort the elements of the first array of scalar values in descending order from, for example, 1 to 2156.

[00159] In block 2412, convert the associated indices into a sequence. For example, algorithm module 1856 can convert the associated indices into a string or dataset sort order instead of the scalar value.

[00160] In block 2414, convert the string to a second array with a scalar value for each unique index which is the distance (difference in position in array rather than measurement in mm) from the primary index from the start of the string array. For example, algorithm module 1856 can convert the string into a second array with a scalar value for each unique index which is the primary index position from the start of the string array with 2165 elements currently, but it can be extensible depending on the data for each individual.

[00161] With reference to FIG. 25, in certain implementations, deep learning can be performed on the complete dataset to establish the minimum dataset needed to accurately perform facial recognition. The result of deep learning can be a reduced dataset of binary resources that is a fraction of the total dataset (eg 1%, 2%, 5%, 10%, 20%, 50%).

Each data element in this dataset can be defined by a sub-array of a specific size and location and a single value within the complete LBP base matrix. Each data element can be assigned a unique numeric identifier. For the application process, this static dataset is used for each applicant. As each requester continues to check their faces, verification data can be collected and, through deep learning, a new set of binary resources can be created for each requester. The introduction of the sequence in the verification algorithm can allow the introduction of this individualized set of binary resources and still allow the comparison between different sets. The matching of two sequences can be achieved by calculating the difference (for example, the difference in the position of a data set) of the two matrices. The larger the sum, the lower the quality of the match. A perfect combination of objects can have a sum of substantially zero.

[00162] Still referring to FIG. 25, a table 2500 illustrates an example of the verification calculations. The first step in creating the sequence can be sorting the scalar array by the magnitude of the scalar value. Sorting can be performed in descending order. The unique identification of the binary resource with the highest value is at the top of the ranking.

Sorting this way, the binary features that provide the greatest uniqueness (highest scalar magnitude) are at the top of the array. After sorting, the new sorting position is transferred to the third matrix and the location in the new matrix is based on the unique identification of the binary resource. The binary resource's unique identifier is 2,000 when it is sorted and goes to position 1. In the third array, position 2,000 will receive a value of 1. This array can be a unique string for the object and is the new basis for verification and /or object match.

[00163] The sequence matching algorithm determines the quality of the match by the distance of each unique binary resource ID from the beginning of the sequence. The distance is the index value of the unique identification of the binary resource within the sequence array where the index values are sequentially assigned. To efficiently complete this correspondence calculation, another matrix is created containing the distance of each binary feature from the beginning of the sequence. For example, after sorting, the binary resource with the unique identification of one is in position ten of the sorted sequence. In the new array, which is indexed by zero, position one will have the value ten. By structuring the matrix in this way, the algorithm for performing a match between two of these matrices is the same as the original algorithm. The absolute value of the difference of values with the match index is aggregated to create a single value for the matrix match. A lower value equals a closer match.

A perfect match can be substantially zero value.

[00164] With the introduction of the sequence array, the matching algorithm can be extensible. The sequence matrix is the second matrix and the scalar matrix is the first (normalized to block size). Extensible might mean that the fixed list of binary resources can now be expanded and contracted as needed to optimize the object matching process while improving the integrity of the matching process. In this implementation, the sequence matrix can be dynamically adjustable based on two separate deep learning functions based on the environment of the edge device and the specific individual face being authorized. The first function may seek to continually optimize the default list of binary resources applied to some or all object matching attempts before developing an individual list of binary resources. The second function can aggregate image data into individual objects and over time develop an optimized list of binary features for the face. Since the individualized list of binary resources exceeds the matching performance of the default list of binary resources by trying to match parallel objects in the input object image data, the default list of binary resources can be replaced by the individualized list of binary resources for that object in live object matching functions.

[00165] In certain implementations, the deep learning mechanism for both processes can receive sample data in parallel with the live object/face matching functions. As image data is received, two lists of binary resources can be applied to it. The first list can be the standard or individualized list of reduced binary resources (eg three most important blocks out of 100 in 10 x 10 macroblock). The application's output from this list can be used in the live mailing process. The second list can be the complete list of binary resources, comprising all possible macroblock sizes (for example, all 100 within the 10 x 10 macroblock), applied to the possible image positions for each ULBP. The complete list of binary resources can be, for example, twenty times the size of the standard or individualized list of binary resources. The deep learning engine can receive a sequence derived from the complete list of binary resources for each object image received into the system. In the case of the standard deep learning process list of binary resources, images can be categorized into training and validation sets. An independent pattern set can be created including objects other than the object being learned. In the case of the individualized list of binary resources, each object identified by the pattern or existing individualized list of binary resources will be placed in the training, validation and testing sets. All other object lists can be placed in verification and test sets.

[00166] A deep learning training session for the default object list of binary resources and for each individual list of binary resources can be run for each new active entry of the object detection system.

[00167] Once the eye detection algorithm locates the eyes, the face data array is transferred to the face recognition algorithm. If the deep learning algorithm has developed a specific list of binary resources, an aggregated list of those binary resources for all users in the system can be created and used to generate the sequence.

Otherwise, the sequence can be generated using the standard set of binary resources.

The string can be transferred to the verification algorithm, where the matching process can determine the identity of the string.

With the identity established, a complete list of binary resources

(all sub-matrices with full LBP histograms) can be generated and transferred to this identity dataset in the deep learning algorithm.

The deep learning algorithm can process the dataset available for that identity (the set can continue to grow with each check) and generate a revised optimal dataset for that identity.

The optimized dataset can be converted to a string and used in the next verification process for that user.

At the same time, the aggregated set of binary features can be updated, if necessary, to convert the next face matrix.

Although the verification process takes place within the local cluster of devices, the deep learning algorithm can take place in both the local and global cluster as a background task.

[00168] As the results of the training session refine the list of binary resources, the respective list can be updated in the live verification/matching process. The result can be a matching system capable of automatically adjusting to the general population of objects and the unique and specific characteristics of each object. The evolution of object data will enable large-scale object matching solutions with more than

100,000 objects capable of the same precision as a small solution (< 1,000).

[00169] Referring now to FIG. 26, an example of the 2600 deep learning process can rely on feedback loops and machine learning to refine the identification process.

[00170] In block 2602, obtain an image matrix.

For example, processor 1802 and/or communication module 1825 can obtain the heat map of face 1730.

[00171] In block 2604, determine if face specific macroblock/ULBP is available. For example, algorithm module 1856 can determine if face specific macroblock/ULBP is available.

[00172] At block 2606, if custom list of binary resources is not available, convert image to generic face detection list of binary resources. For example, algorithm module 1856 can convert image into generic/pattern face detection list of binary resources if custom list is not available.

[00173] In block 2608, get the face detection pattern sequence. For example, algorithm module 1856 can obtain the face detection pattern sequence if the macroblock is not available.

[00174] At block 2610, if custom list of binary resources is available, convert the image to a face detection specific list of binary resources. For example, algorithm module 1856 can convert image into face detection specific list of binary resources if custom list of binary resources (may be in original 2156 or it may be all new based on that individual (eg 10 10 x Additional 10s that are more distinctive for that particular individual) is available.

[00175] In block 2612, perform face detection.

For example, algorithm module 1856 can perform face detection.

[00176] In block 2614, perform eye detection and location. For example, algorithm module 1856 can perform eye detection and location.

[00177] In block 2616, convert the image into a complete list of binary resources. For example, algorithm module 1856 can convert the image into a complete list of binary resources (not just 57, but all 100 of 10 x 10 macroblocks and include all values). Tensor flow is the framework of the deep learning engine provided by GoogleTM. The equation and algorithm for TensorFlow are well known in the art.

[00178] In block 2618, develop face detection list of binary resources using TensorFlow deep learning. For example, algorithm module 1856 can develop face detection list of binary resources using TensorFlow deep learning.

[00179] In block 2620, refined feedback face detection sequence. For example, algorithm module 1856 can feed back the refined face detection sequence. To refine the feature list, the faceplate 1404 can take the entire feature set and collect all the data for the face and derive a new refined set that replaces the default set. The calculation of the distance between the position of the binary feature within two sets can be used as a metric. Each feature would always be in the same position. As the face data gets better, the face-to-face comparison gets smaller and the difference between this face and all the others gets bigger.

[00180] Turning now to FIG. 27, the example of the deep learning process 2600 shown in FIG. 26 can rely on feedback loops and machine learning to refine the identification process. Instead of centralizing the process of deep learning data that is subject to hacking if stored on a central server (single TensorFlow application running on a single central server), particularly in the cloud, the present disclosure can use TensorFlow in a distributed architecture to assess on each reader with privacy protected on all individual devices, rather than stored on a central server.

[00181] In block 2614, perform eye detection and location. For example, algorithm module 1856 can perform eye detection and location.

[00182] In block 2702, get a specific face detection list of binary resources. For example, algorithm module 1856 can obtain a specific face detection list of binary resources if macroblock is available.

[00183] In block 2704, get the standard facial recognition list of binary resources. For example, algorithm module 1856 can convert image to generic face detection list of binary resources if the specific list of binary resources for that individual is not available.

[00184] At block 2706, determine whether the face-specific list of binary features is available.

For example, algorithm module 1856 can determine whether the face-specific list of binary resources is available. The custom list of binary resources can include the original 2156 or it can be all new based on that individual (eg additional 10 10 x 10s that are more distinct for that specific individual)

[00185] In block 2708, if face specific list of binary features is not available, convert image to generic face list of binary features.

For example, algorithm module 1856 can convert image to generic face list of binary resources if face specific list of binary resources is not available.

[00186] At block 2710, if face specific list of binary features is available, convert the image to a face specific list of binary features. For example, algorithm module 1856 can convert image to face specific list of binary features if face specific list of binary features is available.

[00187] In block 2712, convert the list of binary resources into sequences. For example, the algorithm module

1856 can convert list of binary resources into strings.

[00188] In block 2714, perform the check. For example, algorithm module 1856 can perform verification.

[00189] At block 2716, convert the image into a complete list of binary resources. For example, algorithm module 1856 can convert the image into a complete list of binary resources.

[00190] In block 2718, develop face-specific list of binary features using TensorFlow deep learning. For example, algorithm module 1856 can develop a face-specific list of binary features using TensorFlow deep learning (not just 57 as is typical with LBP, but all 100 of 10 x 10 macroblocks and include all values). TensorFlow is the framework for the deep learning engine provided by GoogleTM. The equation and algorithm for TensorFlow are well known in the art.

[00191] In some examples, port 1402 and/or faceplate 1404 may include proactive algorithms to identify and reduce "consistency collisions" during the verification process. A "consistency collision" can occur when one or more applicants have unique facial data that causes an error (ie, false acceptance or false rejection) in verification. Port 1402 and/or faceplate 1404 can proactively check for potential "consistency collisions" using two methods of the present disclosure described below.

[00192] The 'Layered Reinforcement' algorithm can allow the requester to create a binary tree structure for the facial data. Binary data can create a hierarchy of data points (for example, LBP sequence) based on the uniqueness of the data points. The greater the singularity, the higher up the tree. Each verification transaction can result in the transfer of the binary facial data from faceplate 1404 to port 1402. When port 1402 receives the binary facial data, it can start to check the uniqueness of the binary tree structure of the new data against the data of the other 1450 stored applicants. When it identifies a potential collision, port 1402 can notify the faceplate 1404 to escalate the verification transactions between the two identified requesters 1450 to port 1404. When port 1404 receives an escalated transaction, it can perform two advanced algorithms (ie , Proactive Identification Collision Algorithm and Time Domain Trends Algorithm) in the data. The binary tree can have more distinct features and establish a hierarchy based on the majority and minority of features. The prioritized tree can reduce parsing based on 20 instead of 2165, saving time and computation/processing. Based on the tree's branches, more distinct characteristics may allow faceplate 1404 to determine whether or not this is the person, rather than going through all of the 2165 characteristics. Boundaries can be empirical and derived from testing.

[00193] The first advanced algorithm, the Proactive Collision Identification Algorithm, can take data from the binary tree and analyze the facial data based on its location in the binary tree. Binary tree data can be weighted. If weighting the binary data does not produce sufficient differentiation of the data, port 1402 can extend the verification process to find suitable differencing data.

[00194] In some implementations, the facial features of the requesters may change continuously. If the biometric data for a face, such as face 1730 of requester 1450, remains static, face identification can ultimately result in false rejections, requiring face 1730 re-enrollment.

Dynamic adjustments to the biometric data can be continuously applied and the algorithms can determine that the proposed change is not introducing a different face to the biometric data, resulting in a false acceptance on face 17303. With each iteration of the biometric data, the algorithm must determine the amount of biometric data retained to ensure the applicant's next successful identification 1450 and the amount to be changed to ensure the applicant's long-term identification

1450. The determination can be made by a combination of a time domain analysis in which changes are regressed to ensure linearity and facial regional analysis to determine whether the area of change is rational.

[00195] The use of this dynamic data capability leads to the second algorithm, the Time Domain Trends Algorithm. Port 1402 can maintain requester 1450 binary data history and perform time-domain based analysis of the data to assess which resources are changing when requester 1450 binary data history is compared to face 1730 and speed with the which these features are changing. Port 1402 can track the evolution of the face and uses this data to extrapolate and reinforce the unique differences identified in face 1730 of the requester 1450 to establish and further emphasize key features or basic features and assess the rate of change in these key features for the face 1730.

Gate 1402 then determines using time domain based analysis whether these changes are occurring over a period of time to suggest natural changes in face 1730, artificial changes in face 1730 that warrant further analysis, or a mismatch of data history 1450 requester binaries with face

1730. Port 1402 can identify the unique differences and verify that the differences identified in the verification request are consistent with the trend over time. In some implementations, port 1402 can take into account the lifestyle and daily routines of the requester 1450 in the Time Domain Trend Algorithm. For example, if applicant 1450 enjoys outdoor activities, the Algorithm may account for increased tanning during the summer season. The Proactive Collision Identification Algorithm combined with the Time Domain Trending Algorithm allows the method of this modality of the invention to maintain its high performance in a large population '1:N' solution.

[00196] Turning now to FIG. 28, referring to the above figures, in some implementations a method 2800 of constructing a biometric template may be performed by the faceplate 1404.

[00197] In block 2802, emitting a non-visible incident light. For example, the light source 1704 can emit non-visible incident light. Non-visible light can include near IR or UV. In some implementations, the light source 1704 can emit visible light.

[00198] In block 2804, detecting a detected non-visible light, wherein the detected non-visible light includes a reflected non-visible light and a radiated non-visible light.

For example, optical sensor 1702 can detect a detected non-visible light, where the detected non-visible light includes a reflected non-visible light and a radiated non-visible light. In some implementations, optical sensor 1702 can detect IR light reflected from face 1730 of requester 1450 and/or IR light radiated due to heat from requester 1450.

[00199] In block 2806, generate a heat map including a plurality of sampling points with a plurality of characteristic values associated with the detected non-visible light. For example, algorithm module 1856 may generate a heatmap, such as heatmap 2000, including a plurality of sampling points with a plurality of characteristic values associated with the detected non-visible light.

[00200] In block 2808, identify one or more macroblocks each including a subset of the plurality of sampling points. For example, algorithm module 1856 may identify one or more macroblocks,

each including a subset of the plurality of sampling points. The one or more macroblocks can be a 10x10 macroblock, a 15x15 macroblock, a 20x20 macroblock, a 30x30 macroblock, a 35x35 macroblock, and a 40x40 macroblock. Other sizes are possible. In some implementations, faceplate 1404 may identify 2165 macroblocks with 2165 associated.

[00201] In block 2810, select a local default value. For example, algorithm module 1856 may select a local binary default value of 20. In other examples, algorithm module 1856 may select a local ternary default value.

[00202] In block 2812, calculate a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks. For example, algorithm module 1856 may compute a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks.

[00203] In block 2814, generating a first matrix including a plurality of weighted values, calculating the plurality of weighted values based on the number of occurrences of the local standard value and corresponding sizes of one or more macroblocks. For example, algorithm module 1856 may generate a first matrix including the weighted values shown in table 2300 (ie, [0.30 0.54 0.15 0.42 0.62 0.46 0.53]).

[00204] In block 2816, assign a unique index to each of the plurality of weighted values. For example, algorithm module 1856 may assign unique indices (i.e., 1, 2, 3, 4, 5, 6, and 7) to each of the plurality of weighted values shown in table 2300. Unique index 1 is assigned to weighted value from 0.30, 2 to 0.54, 3 to 0.15..., etc.

[00205] In block 2818, generate a second array of the unique index sorting the plurality of weighted values. For example, algorithm module 1856 may generate a second unique index array by sorting the plurality of weighted values, such as the sequence 5, 2, 7, 6, 4, 1, 3 in table 2300. In some implementations, the second array /string can indicate a ranking of weighted values from highest to lowest. The first number in the sequence is 5 because the weighted value (that is, 0.62) associated with the unique index of 5 is the highest among the elements of the first array.

[00206] In block 2820, generate a third matrix including a plurality of classification distances.

For example, algorithm module 1856 may generate a third matrix including a plurality of rank distances, such as the matrix stored [5 1 6 4 0 3 2] in table 2300. A rank distance may indicate a numerical difference between ranks of the highest weighted value (eg 0.62 - rank 1) and the current weighted value (eg 0.30 - rank 6). Therefore, the classification distance between 0.62 and 0.30 can be 5 (ie 6-1).

[00207] It will be recognized that various implementations of the above-disclosed features and functions and others, or alternatives or varieties thereof, may desirably be combined in many other different systems or applications by a person skilled in the art.

Furthermore, various presently unforeseen or unforeseen alternatives, modifications, variations or improvements may subsequently be made by those skilled in the art who are also intended to be covered by the following claims.

[00208] The detailed description set out above in connection with the accompanying drawings describes examples and does not represent the only examples that can be implemented or that fall within the scope of the claims. The term "example", when used in this description, means "serving as an example, instance or illustration" and not "preferred" or "advantaged over other examples". The detailed description includes specific details for the purpose of providing an understanding of the techniques described. These techniques, however, can be practiced without these specific details. For example, changes can be made to the function and arrangement of the elements discussed without departing from the scope of disclosure. In addition, multiple examples can omit, replace, or add multiple procedures or components, as appropriate. For example, the described methods can be performed in a different order than described and various steps can be added, omitted or combined. Also, the features described in relation to some examples can be combined in other examples. In some cases, well-known structures and apparatus are shown in block diagram form in order to avoid obscuring the concepts of the examples described.

[00209] Information and signals can be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols and chips that may be referenced throughout the description above may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, executable code by computer or instructions stored on a computer-readable medium, or any combination thereof.

[00210] The various illustrative blocks and components described in connection with the disclosure in this document may be implemented or executed with a specially programmed device, such as, but not limited to, a processor, a digital signal processor (DSP), an ASIC , an FPGA or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination thereof designed to perform the functions described herein. A specially programmed processor can be a microprocessor, but alternatively the processor can be any conventional processor, controller, microcontroller or state machine. A specially programmed processor can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors together with a DSP core, or any other configuration.

[00211] The functions described here can be implemented in hardware, software executed by a processor, firmware or any combination thereof. If implemented in software executed by a processor, functions can be stored or transmitted as one or more instructions or code in a non-transient computer-readable medium. Other examples and implementations are within the scope and essence of the disclosure and accompanying claims. For example, due to the nature of software, the functions described above may be implemented using software executed by a specially programmed processor, hardware, firmware, cabling, or combinations of any of these. Features that implement functions can also be physically located in various positions, including being distributed so that portions of functions are implemented in different physical locations. Furthermore, as used in this document, including the claims, "or" as used in a list of items preceded by "at least one of" indicates a disjunctive list such that, for example, a list of "at least one of A, B or C” means A or B or C or AB or AC or BC or ABC (ie A and B and C).

[00212] Computer readable media includes computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another. A storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer. By way of example, and not limitation, computer readable media may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage,

magnetic disk storage or other magnetic storage devices, or any other medium that can be used to transport or store desired program code media in the form of instructions or data structures and that can be accessed by a general purpose computer or computer. special use, or a general purpose or special purpose processor. In addition, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wireless technologies such as infrared, radio and microwave , then coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the medium definition.

Disk and floppy disk, as used herein, include compact disk (CD), laser disk, optical disk, digital versatile disk (DVD), floppy disk, and Blu-ray disk where floppy disks generally reproduce data magnetically, while disks reproduce data optically with lasers . Combinations of the above items are also included in the scope of computer readable media.

Claims (20)

1. Faceplate for biometric identification characterized by the fact that it comprises: a light source configured to emit non-visible incident light; an optical sensor configured to detect a detected non-visible light, wherein the detected non-visible light includes a reflected non-visible light and a radiated non-visible light; one or more processors operatively coupled to the light source and the optical sensor, the one or more processors are configured to build a biometric model of a requester requesting access to an entry point by: generating a heat map including a plurality of points of sampling having a plurality of characteristic values associated with the detected non-visible light, identifying one or more macroblocks each including a subset of the plurality of sampling points, selecting a local pattern value, calculating a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks; generating a first matrix including a plurality of weighted values, calculating the plurality of weighted values based on the number of occurrences of the local pattern value and corresponding sizes of one or more macroblocks; assigning a unique index to each of the plurality of weighted values; generating a second matrix of the unique index sorting the plurality of weighted values; and generating a third matrix including a plurality of classification distances. 2. Front plate, according to claim 1, characterized in that it further comprises: a monitor configured to display an image of a requester's face captured by the optical sensor. 3. Front plate, according to claim 1, characterized in that: the reflected non-visible light is generated from the incident non-visible light reflected on a face of the applicant; and radiated non-visible light is generated from the applicant's heat. 4. Front plate, according to claim 1, characterized in that the non-visible incident light is close to infrared or ultraviolet light. 5. Front plate, according to claim 1, characterized in that it further comprises a digitizer configured to: transmit an interrogation signal to a proximity card of the applicant, and receive a response signal from the proximity card including an identification sequence . 6. Front plate, according to claim 1, characterized in that it additionally comprises: a memory that stores a plurality of biometric models of authorized personnel; and wherein one or more processors are further configured to: compare the applicant's biometric template with the plurality of authorized personnel biometric templates, and generate a positive match signal in response to identifying a match between the applicant's biometric template and a of the plurality of biometric templates of authorized personnel. 7. Front plate, according to claim 6, characterized in that it further comprises a modem configured to transmit the positive match signal to an access port to authorize the requester to access the entry point. 8. Front plate, according to claim 1, characterized in that the biometric template includes a facial recognition template. 9. Faceplate according to claim 1, characterized in that the local pattern value includes a local binary pattern value or a local ternary pattern value. 10. Biometric identification method characterized by the fact that it comprises: emitting a non-visible incident light; detecting a detected non-visible light, wherein the detected non-visible light includes a reflected non-visible light and a radiated non-visible light; generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light; identifying one or more macroblocks each including a subset of the plurality of sampling points; selecting a local pattern value, calculating a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks; generating a first matrix including a plurality of weighted values, calculating the plurality of weighted values based on the number of occurrences of the local pattern value and corresponding sizes of one or more macroblocks; assigning a unique index to each of the plurality of weighted values; generating a second matrix of the unique index sorting the plurality of weighted values; and generating a third matrix including a plurality of classification distances. 11. Method according to claim 10, characterized in that it further comprises: displaying an image of a face of an access requester captured by the optical sensor. 12. Method according to claim 10, characterized in that: the reflected non-visible light is generated from the incident non-visible light reflected on a face of the requester; and radiated non-visible light is generated from the applicant's heat. 13. Method according to claim 10, characterized in that the non-visible incident light is close to infrared light or ultraviolet light. 14. Method according to claim 10, characterized in that it further comprises: transmitting an interrogation signal to an applicant's proximity card, and receiving a response signal from the proximity card including an identification string. 15. Method according to claim 10, characterized in that it further comprises: storing a plurality of biometric templates of authorized personnel; compare the applicant's biometric template with the plurality of authorized personnel biometric templates; and generating a positive match signal in response to identifying a match between the applicant's biometric template and one of the plurality of authorized personnel biometric templates. 16. Method according to claim 15, characterized in that it further comprises transmitting the positive match signal to an access door to authorize the requester's access to the entry point. 17. Method according to claim 10, characterized in that the biometric template includes a facial recognition template. 18. Method according to claim 10, characterized in that the local pattern value includes a local binary pattern value or a local ternary pattern value. 19. The method of claim 10, wherein the one or more macroblocks include at least one of a 10 x 10 macroblock, a 15 x 15 macroblock, a 20 x 20 macroblock, a 30 x 30 macroblock, a 35 x 35 macroblock, and a 40 x 40 macroblock. 20. Infrastructure characterized by the fact that it comprises: an entry point with controlled access; and a front panel configured to: emit a non-visible incident light on the face of an applicant; detecting a non-visible light detected from the face of the requester, wherein the non-visible light detected includes a non-visible reflected light and a non-visible radiated light; generating a biometric model of the applicant by: generating a heat map including a plurality of sampling points having a plurality of characteristic values associated with the detected non-visible light; identifying one or more macroblocks each including a subset of the plurality of sampling points; select a local default value, calculating a number of occurrences of the local pattern value within each subset of the plurality of sampling points for each of the one or more macroblocks; generating a first matrix including a plurality of weighted values, calculating the plurality of weighted values based on the number of occurrences of the local pattern value and corresponding sizes of one or more macroblocks; assigning a unique index to each of the plurality of weighted values; generating a second matrix of the unique index sorting the plurality of weighted values; and generating a third matrix including a plurality of classification distances; store a plurality of biometric templates of authorized personnel; compare the applicant's biometric template with the plurality of authorized personnel biometric templates; generating a positive match signal in response to identifying a match between the applicant's biometric template and one of the plurality of authorized personnel biometric templates; and transmitting the positive match signal to an access port to grant the requester access to the entry point.