Systems and methods for secure biometric identification using recorded pressure

Abstract

This article describes a system and method for secure biometric identification using dynamic pressure sensing that is easy to use and intuitive. Accurate identification is accomplished by using a set of finely spaced analog sensors that measure and output a dynamic pressure profile, which is then evaluated based on data from a trained model. The model includes multiple personal biometric characteristics that can be used to uniquely identify a person, for example, for authentication purposes, such as in conjunction with electronic business transactions, Internet of Things (IoT) devices, automotive devices, identity and access management (IAM), or robotics or high-function touch sensing devices to grant access to sensitive confidential information.

Description

[0001] Cross-references to related patent applications

[0002] This application claims the benefit of priority under 35 U.S.C. § 119(e) to co-pending and commonly assigned U.S. Patent Application No. 62 / 730,948, filed on September 13, 2018, entitled “SYSTEMS AND METHODS FOR SECURE BIOMETRIC IDENTIFICATION USING RECORDED PRESSURE”, inventors Hong Chang and Alex Cheuk Kwong Lee, the entire contents of which are incorporated herein by reference. Each reference cited in that patent document is incorporated herein by reference in its entirety. Technical Field

[0003] The present disclosure generally relates to systems and methods for personal identification based on human-computer interaction. More specifically, the present disclosure relates to biometric systems and methods that use dynamic pressure sensing to uniquely identify a user. Background Art

[0004] Current biometric methods for personal identification use fingerprint data, retinal scans, facial recognition methods, and the like. All of these identification methods have the following in common: they operate on static biometric data, i.e., each of the biometric features is unique to each individual and serves as the basis for identification data that, for practical purposes, does not change over time. An obvious disadvantage of static biometric methods for personal identification is that such methods can be relatively easily circumvented by a motivated attacker or identity thief by copying and then reproducing the identification data to mimic the presence of a user with unique features. For example, by presenting a facial recognition system with a color copy of a picture of the person to be identified, an impostor can successfully trick the security system into misidentifying the photo as that person, thereby defeating the security system and gaining access to confidential information.

[0005] Therefore, what are needed are dynamic systems and methods that are virtually impossible to copy and are therefore inherently more secure for protecting confidential information. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Reference will be made to embodiments of the present invention, examples of which may be shown in the accompanying drawings. These drawings are illustrative only and non-restrictive. Although the present invention is generally described in the context of these embodiments, it should be understood that it is not intended to limit the scope of the present invention to these specific embodiments.

[0007] FIG. 1 is a simplified block diagram of a secure biometric identification system according to various embodiments of the present disclosure.

[0008] Figure 2A An exemplary input buffer list according to various embodiments of the present disclosure is illustrated.

[0009] Figure 2B An exemplary database buffer list according to various embodiments of the present disclosure is illustrated.

[0010] Figure 3 An exemplary hybrid buffer list according to various embodiments of the present disclosure is illustrated.

[0011] Figure 4 Exemplary data structures according to various embodiments of the present disclosure are illustrated.

[0012] Figure 5 Exemplary user accounts according to various embodiments of the present disclosure are illustrated.

[0013] Figure 6 is a flow chart of an illustrative process for secure biometric identification using recorded pressure according to various embodiments of the present disclosure.

[0014] Figure 7 is a flow chart of an illustrative process for merging data according to various embodiments of the present disclosure. DETAILED DESCRIPTION

[0015] In the following description, for the purpose of explanation, specific details are set forth in order to provide an understanding of the present invention. However, it will be apparent to those skilled in the art that the present invention can be practiced without these details. In addition, those skilled in the art will recognize that the embodiments of the present invention described below can be implemented in a variety of ways, such as a process, an apparatus, a system, a device, or a method on a tangible computer readable medium.

[0016] The components or modules shown in the figures are illustrations of exemplary embodiments of the present invention and are intended to avoid obscuring the present invention. It should also be understood that throughout this discussion, components may be described as separate functional units, which may include subunits. Those skilled in the art will recognize that various components or parts thereof may be divided into separate components or may be integrated together, including being integrated into a single system or component. It should be noted that the functions or operations discussed herein may be implemented as components. Components may be implemented with software, hardware, or a combination thereof.

[0017] In addition, the connections between components or systems within the drawings are not intended to be limited to direct connections. Rather, data between these components may be modified, reformatted, or otherwise altered by intermediate components. Moreover, additional or fewer connections may be used. It should also be noted that the terms "coupled," "connected," or "communicatively coupled" are understood to include direct connections, indirect connections through one or more intermediate devices, and wireless connections.

[0018] References in the specification to "one embodiment," "preferred embodiment," "embodiment," or "embodiments" mean that a particular feature, structure, characteristic, or function described in conjunction with the embodiment is included in at least one embodiment of the invention, and may be in more than one embodiment. Moreover, the appearances of the above phrases in various places in the specification do not necessarily all refer to the same embodiment or embodiments.

[0019] The use of certain terms in various places in the specification is for illustrative purposes and should not be construed as limiting. A service, function, or resource is not limited to a single service, function, or resource; use of these terms may refer to groupings of related services, functions, or resources, which may be distributed or aggregated. In addition, the use of memory, database, repository, data storage device, table, hardware, etc. may be used herein to refer to one or more system components in which information may be entered or otherwise recorded.

[0020] In this document, "sensor array" and "grid" may be used interchangeably. Similarly, "pressure" and "force" may be used interchangeably because, as will be appreciated by those skilled in the art, pressure is defined as a force applied to a given surface area. Thus, the term "pressure data" may refer to data indicative of pressure or a quantity (e.g., a physical quantity of force) from which pressure may be derived. The terms "pressure sensor" and "pixel" refer to a device for measuring, directly or indirectly, pressure or force at one or more locations, i.e., a quantity from which pressure data may be derived.

[0021] Figure 1 1 is a simplified block diagram of a secure biometric identification system according to various embodiments of the present disclosure. The system 100 includes a GUI / API 104, an error handling module 106, a security handling module 108, a machine training module 110, a data analysis module 112, a device driver 114, a client / server API 116, and a pressure recorder 120. It should be understood that the modules described herein can be implemented by a state machine using wired logic or in software using a processing unit such as a microcontroller to execute instructions.

[0022] Examples of pressure sensors and pressure recorders 120 that may benefit from the teachings of the present disclosure are the systems and methods disclosed in U.S. Provisional Patent Application Serial No. 62 / 729738, filed on September 11, 2018, entitled “Pressure Recording Systems and Methods for Biometric Identification,” inventors Alex Cheuk Kwong Lee and Hong Chang, the entire contents of which are incorporated herein by reference for all purposes.

[0023] In an embodiment, the input data 122 may include data provided by a multi-dimensional array of pressure sensors, for example, spatial and temporal pressure data, such that the data captured by the pressure recorder 120 may represent a spatial pressure distribution that may vary over time. In an embodiment, the input data 122 may be based on fused data from any number of different types of sensors. The temporal variation of the pressure data may be continuous, or, for example, according to discrete periodic time intervals. In an embodiment, to ensure accuracy, the input data 122 may include pressure measurement data stored as pixel buffer values ​​in a time series.

[0024] Note that the pressure recorder 120 may be a high-resolution pressure recorder 120 that may accurately capture, for example, at dynamically changing measurement rates, the nuances of gestures or motions applied even unconsciously when a user interacts with the pressure sensor 120, for example, using a finger or another body part. In an embodiment, the data analysis module 112 may apply relational analysis to infer or derive data 122 from the movements recorded by the user of the pressure recorder 120, such as the rolling of the user's thumb while applying pressure to the pressure touch sensor.

[0025] In an embodiment, the data 122 can be used to generate a record (e.g., frame by frame) representing spatially distributed pressure data, such that each frame can be assigned multiple (e.g., different) patterns or sub-patterns associated with multiple pressure regions. In an embodiment, the input data 122 can be pre-processed in a manner that facilitates comparison of the patterns to each other, such as being classified, grouped, and processed. In an embodiment, the pre-processing includes fusing data from two or more sources that may include additional pressure-related data. Similarly, data from multiple regions can be processed before being used as training data and stored in a database.

[0026] In an embodiment, the input data 122 may be used to learn a set of (sub) patterns, which may be associated with, for example, physical parameters such as speed, range, etc., and may be used to identify, for example, multiple characteristics within a pattern and / or characteristics that are common to two or more patterns. In an embodiment, patterns may be selected from a (e.g., random) sample and stored in a database, e.g., in the data analysis module 112. The learned characteristics may then be used to identify the user that caused the generation of the pattern(s).

[0027] In an embodiment, once the input data 122 has been received, machine learning may be applied thereto in order to identify patterns and other characteristics therein to generate groups and subgroups. In an embodiment, a weighting factor may be assigned to each group and even each pixel, for example, based on a set of predetermined criteria. The machine training module 110 may adjust the weighting factors and other parameters over time, for example, to accommodate natural variations in user pressure and speed responses. The machine training module 110 may apply and update the training method at various times, for example, based on gestures or movements that have been entered into the database after being recognized as valid.

[0028] Once the temporal and spatial characteristics of the input data 122 are successfully mapped to known gestures in the database, the user can be identified and the identification process can end. In an embodiment, the machine training module 110 takes into account tolerances, such as pressure variations caused by temperature and other environmental parameters, which may otherwise negatively affect sensor accuracy, and therefore the accuracy of the pressure recorder 120. It should be understood that tolerances can also be user-dependent, for example due to factors such as the user's physical condition that may vary over time - whether within a day or within a decade.

[0029] The learning of patterns and characteristics may be performed by the machine training module 110, which receives the input data 122 as training data. In an embodiment, for example, in a training phase, after an initialization process that may involve prompting the user to provide a username via the user interface 104 (not shown), the user may be prompted to apply pressure on the pressure sensor using his thumb, finger, or any other body part.

[0030] In an embodiment, for example, the user may be prompted to repeatedly press the sensor multiple times with varying forces, speeds, and positions, such as by rotating or rolling a thumb, resulting in different angles and / or sequences, where the pressure sensor captures user-generated pressure-related data (e.g., temporal and spatial information) as well as environmental parameters (e.g., temperature, etc.), which may then be pre-processed and stored in a database, such as within the pressure recorder 120.

[0031] In an embodiment, the machine training module 110 may use an averaging technique, for example, averaging the pressure data 122 over a period of time, an area, or both, for example to determine an average location (similar to the center of gravity) of the pressure distribution in each frame or frame area. Activity levels may then be defined and assigned to one or more regions within each frame. For example, the pressure data within a region may be averaged to determine the highest number or active pixels, i.e., pixels with values ​​above a particular threshold. In an embodiment, a pixel may be defined as active based on the number of less active neighboring pixels within a given frame or based on data continuity. In an embodiment, multiple active pixels may be clustered into subgroups or regions that may be collectively classified as "active regions." In an embodiment, regions may be combined to form a data structure representing a user profile, which may then be used to identify the user. In an embodiment, a set of the most active pixels may be used to map one orientation to another to account for the tilt of the reference frame.

[0032] In an embodiment, the database may store historical data that may be used as contextual data for interpreting or evaluating the significance of sensor data from the pressure recorder 120. For example, knowing the fact that the user is a child or an elderly person when generating a user profile may help to more accurately interpret the data 122, for example by taking into account that certain categories of users typically apply less pressure to the sensor when so prompted. Similarly, a child with underdeveloped motor skills may appear less consistent and therefore generate responses with greater variation when asked to press the sensor five times in a row compared to an adult.

[0033] Thus, in embodiments, historical data may be used as a factor in determining confidence intervals that may define what is considered reliable data. Conversely, in embodiments, receiving identical or highly accurate data may be used as an indicator that an attacker is somehow using repeated or replicated real-time pressure data to simulate individual pressure characteristics of a certain user, based on the assumption that no one can reproduce two identical gestures by applying pressure to a surface. Thus, in embodiments, a confidence interval may be defined by a volatility index (VIX), i.e., a range or value of expected accuracy that may take into account sample-to-sample differences. The machine training module 110 may calculate the VIX as a user-specific confidence range, for example, based on differences in user input data during a training phase. In embodiments, to further improve security, the security manipulation module 108 may, among other things, manipulate user access rights using any security model known in the art and adjust confidence levels based on expected environmental and user-related tolerances.

[0034] In an embodiment, the time series comparison database 602 can be used to generate a more compact hybrid comparison database 606 than can be delivered to the client-side server, where it can be reorganized into one or more non-time-based groups based on the AI-defined characteristics 614. In addition, the format of the database can be selected to be significantly different from the input data 122, so that the input data 122 cannot be imitated to circumvent the security of the system 100. In an embodiment, the characteristics of each data may be based on neighboring data and data collected immediately before or after. In an embodiment, a small random adjustment in the input data 122 can significantly change the data characteristics and thus trigger a recognition failure.

[0035] In an embodiment, input data 122 changes the contents of database buffer list 602, hybrid buffer list 606, and new hybrid data structure 614, so that if an attacker manages to steal information contained in these databases, the static information will become outdated. In an embodiment, authentication does not rely on direct comparison of these databases, further hindering any attacker.

[0036] In an embodiment, to reduce computation time, only the necessary portion of the database may be transferred to the client side. This keeps the memory allocation on the client side reasonably small and reduces bandwidth requirements when transferring data over the Internet. In an embodiment, to reduce the computational burden, multiple frames with the highest number of active pixels may be used for evaluation purposes. In an embodiment, to achieve this, the pressure data within the frames may be summed to obtain the frame with the most active points, i.e. the frame associated with the highest total pressure. This frame may then be used as the basis for generating a unique profile / highest resolution feature.

[0037] Figure 2A An exemplary input buffer list according to various embodiments of the present disclosure is illustrated. Figure 2B An exemplary database buffer list according to various embodiments of the present disclosure is illustrated. Figure 3 An exemplary hybrid buffer list according to various embodiments of the present disclosure is illustrated. Figure 4 Exemplary data structures according to various embodiments of the present disclosure are illustrated. Figure 5 Exemplary user accounts according to various embodiments of the present disclosure are illustrated.

[0038] Figure 66 is a flow chart of an illustrative process for secure biometric identification using recorded pressure according to various embodiments of the present disclosure. In an embodiment, the data conclusion module 604 receives historical pixel pressure values ​​from the database buffer list 204 or pixel values ​​from the input buffer list 201 and calculates the relative relationship of the pixel to its corresponding neighbor based on it. It should be understood that the calculation can be performed for some or all pixels in the buffer. In an embodiment, one or more thresholds are used to filter or classify the pixel data 207 or 203, for example, to find pixel boundaries 305, hot spots 303, intensity data 304, or VIX data 308 for the buffer. Any of the filtered / classified data can be combined into a mixed bag 302. In an embodiment, the steps starting with the calculation can be repeated for each buffer 205 or 202 to output a list 301 of mixed buffers 302, which can include information about pixel boundaries 305, hot spots 303, intensity data 304, and VIX data 308 for each mixed buffer 302. In an embodiment, the hybrid buffer 302 may include an input buffer pointer 306 and a database buffer pointer 307 which may constitute a null pointer.

[0039] In an embodiment, data processor 610 receives input from at least one of input buffer list 201, hybrid buffer list 301, or database buffer list 204. When input is not received from hybrid buffer list 301, then it may be received from input buffer list 201 or database buffer list 204.

[0040] In an embodiment, the data processor uses the list 301 in the data conclusion module 604 to find the key buffer with the largest boundary, the largest hot spot, or the largest intensity in the hybrid buffer 302. In an embodiment, the data processor 610 reconstructs the buffer 302 into, for example, new groups 402, and assigns priorities 404 to these groups, for example, based on calculated values ​​representing defined characteristics. In an embodiment, the group 402 may be divided into sub-groups, each of which undergoes a similar process in the recursive function until a new data structure 401 is output. The new data structure 401 may or may not have a handle point back to the original buffer 205 or 202 input to the function.

[0041] In an embodiment, the data comparison module 616 receives as input at least one of: a hybrid buffer list 301; a new input hybrid buffer list 301 with a handle to the input buffer 202; or a valid comparison value range 504 for a particular user. In an embodiment, the data comparison module 616 aligns the key buffer 403, compares each group 402, multiplies the comparison result by the priority 404, and each group 402 result together, and compares the calculated result with the user's range 504. After the calculated value falls within the range, the process 600 can continue to the next described data merge process.

[0042] Figure 7 is a flow chart of an illustrative process for merging data according to various embodiments of the present disclosure. In an embodiment, data merge process 700 includes receiving input data from new data structure 401 with a handle pointing back to database buffer 205 and new data structure 401 with a handle pointing back to input buffer 202.

[0043] In an embodiment, the data merge process 700 aligns the key buffers 403 and separately merges each new buffer group 402. In an embodiment, if two new data structures 401 do not have the same length within a group 402, the corresponding buffers 205 may be trimmed or padded, and a decision may be made as to which has a stronger characteristic.

[0044] In an embodiment, the process 700 includes mapping each pixel 207 to the historical pressure value 203 in each frame, and recalculating the pixel history value 207, recalculating the database buffer history value 206, and / or recalculating the valid comparison value range 504 for a given user. In an embodiment, the new buffer group 402 can be added together to form an updated new database buffer list 204. In an embodiment, the results can be written to the database buffer list 204.

[0045] Aspects of the present invention may be encoded in instructions on one or more non-transient computer-readable media for one or more processors or processing units to cause the steps to be executed. It should be noted that one or more non-transient computer-readable media should include volatile and non-volatile memory. It should be noted that alternative implementations are possible, including hardware implementations or software / hardware implementations. Application-specific integrated circuits (ASICs), programmable arrays, digital signal processing circuits, etc. can be used to implement the functions of hardware implementations. Therefore, the terms in any claim are intended to cover software and hardware implementations. Similarly, the term "computer-readable medium" as used herein includes software and / or hardware with a program of instructions implemented thereon, or a combination thereof. In view of the substitution of these implementations, it should be understood that the drawings and the accompanying descriptions provide functional information that those skilled in the art will need to write program code (i.e., software) and / or manufacture circuits (i.e., hardware) to perform the desired processing.

[0046] It should be noted that embodiments of the present invention may also relate to computer products with non-transient tangible computer-readable media having computer codes thereon for performing various computer-implemented operations. The media and computer codes may be those specifically designed and constructed for the purposes of the present invention, or they may be of a type known or available to those skilled in the relevant art. Examples of tangible computer-readable media include, but are not limited to: magnetic media such as hard disks; optical media such as CD-ROMs and holographic devices; magneto-optical media; and hardware devices specifically configured to store or store and execute program codes, such as ASICs, programmable logic devices (PLDs), flash memory devices, and ROM and RAM devices. Examples of computer codes include machine codes such as those generated by a compiler, and files containing high-level codes executed by a computer using an interpreter. Embodiments of the present invention may be implemented in whole or in part as machine executable instructions, which may be in program modules executed by a processing device. Examples of program modules include libraries, programs, routines, objects, components, and data structures. In a distributed computing environment, program modules can be physically located in a local, remote, or both setting.

[0047] Those skilled in the art will recognize that no computing system or programming language is critical to practicing the present invention. Those skilled in the art will also recognize that a number of the above elements may be physically and / or functionally separated into submodules or combined together.

[0048] Those skilled in the art will appreciate that the foregoing examples and embodiments are illustrative and not limiting of the scope of the present disclosure. It is intended that all permutations, enhancements, equivalents, combinations and improvements that are apparent to those skilled in the art after reading the specification and studying the drawings are included within the true spirit and scope of the present disclosure. It should also be noted that the elements of any claim may be arranged differently, including having multiple dependencies, configurations and combinations.

Claims

1. A biometric identification method, comprising: receiving input data, the input data comprising spatial and temporal pressure data that has been generated in response to a user applying pressure on the pressure sensor by depressing the pressure sensor, the pressure sensor comprising a plurality of regions, wherein at least some of the spatial and temporal pressure data has been recorded in frames, each frame being assigned an activity level based on a set of active pixels, the activity level being defined by the pressure data, the pressure data having been averaged over a region to determine a highest number of pixels exceeding a threshold, defining a pixel as active based on a number of less active neighboring pixels; using the set of active pixels to map the orientation to another orientation to account for a tilt of a reference frame; using the input data to identify one or more patterns associated with the plurality of regions; associating the one or more patterns with one or more characteristics; and The one or more characteristics are used to identify the user.

2. The biometric feature recognition method according to claim 1, further comprising: comparing the one or more patterns with patterns already obtained in a training phase by a trained model updated based on successfully learned features; calculating a probability that the one or more patterns match the one or more characteristics based on the comparing; and The user is authenticated in response to the probability satisfying a criterion.

3. The biometric feature recognition method according to claim 1, wherein: The input data includes fused data from two or more sensors.

4. The biometric feature recognition method according to claim 1, wherein: The spatial and temporal pressure data have been acquired at varying measurement rates. 5 . The biometric feature recognition method according to claim 1 , further comprising clustering the group of active pixels into active regions.

6. The biometric recognition method according to claim 1 further includes assigning weight factors to elements in the plurality of regions based on a set of predetermined criteria.

7. The biometric recognition method according to claim 6 further includes adjusting the weight factor to take into account at least one of a user-related tolerance or a non-user-related tolerance that causes the change in the pressure.

8. The biometric method of claim 7, further comprising, in response to receiving identical or highly accurate data, using the data as an indication of an attack.

9. The biometric feature recognition method according to claim 1, wherein: Using the input data includes evaluating the input data using context data.

10. The biometric recognition method of claim 1, further comprising evaluating sample-to-sample variation by comparing the sample to a series of expected data.

11. The biometric recognition method according to claim 1 further includes selecting a set of data from among the spatial and temporal pressure data to be used as an orientation reference data set.

12. A biometric feature recognition system, comprising: a pressure recorder comprising a pressure sensor, the pressure sensor comprising a plurality of regions, the pressure sensor receiving input data associated with spatial and temporal pressure data that has been generated in response to a user applying pressure on the pressure sensor, wherein the spatial and temporal pressure data has been recorded in frames, each frame being assigned an activity level based on a set of active pixels in an active region, the activity level being defined by the pressure data, the pressure data having been averaged within the region to determine a highest number of pixels exceeding a threshold, a pixel being defined as active based on a number of less active neighboring pixels; adjusting the spatial orientation using the set of active pixels to account for the tilt of the reference frame; and One or more processors that, in response to receiving the spatial and temporal pressure data, perform steps comprising: identifying one or more patterns associated with the plurality of regions using the trained model updated based on the successfully learned features; associating the one or more patterns with one or more characteristics; and The one or more characteristics are used to identify the user.

13. The biometric recognition system according to claim 12, wherein: The biometric system is used to authenticate or identify the user in conjunction with at least one of: an electronic commerce transaction, an Internet of Things (IoT) device, an automotive device, identity and access management (IAM), or a robotic or high-functioning touch sensing device.

14. The biometric recognition system according to claim 12, further comprising: Weighting factors are assigned to elements in the plurality of regions based on a set of predetermined criteria, and the weighting factors are adjusted to account for at least one of a user-related tolerance or a non-user-related tolerance that causes the change in pressure.

Images