US011281755B2

( 12 ) United States Patent

Chang et al.
( 10) Patent No .: US 11,281,755 B2
(45) Date of Patent : Mar. 22 , 2022
( 54 ) SYSTEMS AND METHODS FOR SECURE ( 56 ) References Cited
BIOMETRIC IDENTIFICATION USING
RECORDED PRESSURE U.S. PATENT DOCUMENTS

( 71 ) Applicants: Hong Chang , Mountain View , CA 7,356,168 B2 4/2008 Tavares

2011/0050394 Al 3/2011 Zhang et al .
(US ) ; Alex C Lee , Mountain View , CA 2014/0157911 A1 6/2014 Sarrafzadeh et al .
( US ) 2015/0282766 A1 10/2015 Cole et al .
( Continued )
( 72 ) Inventors: Hong Chang , Mountain View , CA
(US ) ; Alex C Lee , Mountain View , CA FOREIGN PATENT DOCUMENTS
( US )
CA 2976049 A1 8/2016
( * ) Notice: Subject to any disclaimer, the term of this WO 2011072284 Al 6/2011
patent is extended or adjusted under 35
U.S.C. 154 ( b ) by 181 days . OTHER PUBLICATIONS
( 21 ) Appl. No .: 16 /548,611 International Search Report and Written Opinion dated Nov. 19 ,
2019 , in related International Patent Application No. PCT /US19 /
(22 ) Filed : Aug. 22, 2019 47770 , filed Aug. 22 , 2019 ( 11 pgs ) .
(Continued )
( 65 ) Prior Publication Data
US 2020/0089858 A1
Primary Examiner - Ruiping Li
Mar. 19 , 2020 (74 ) Attorney, Agent, or Firm - North Weber & Baugh
LLP ; Michael North
Related U.S. Application Data ( 57 ) ABSTRACT
( 60 ) Provisional application No. 62/ 730,948 , filed on Sep. Described herein are systems and methods for secure bio
13 , 2018 metric identification using dynamic pressure sensing that are
( 51 ) Int . Cl . convenient and intuitive to use . Accurate identification is
GOOF 21/32 ( 2013.01 )
accomplished by using a set of finely spaced analog sensors
GO6K 9/62 that measure and output a dynamic pressure profile that is
( 2006.01 ) then evaluated based on data from a trained model. The
GOON 20/00 ( 2019.01 ) model comprises a number of personal biometric character
G06F 3/0487 ( 2013.01 ) istics that may be used to uniquely identify a person, e.g. , for
(52 ) U.S. CI . authentication purposes , such as granting access to sensitive ,
CPC G06F 21/32 (2013.01 ) ; G06F 3/0487 confidential information in connection with an electronic
(2013.01 ) ; G06K 986255 ( 2013.01 ) ; GOON commercial transaction, an Internet of Things ( IoT ) device ,
20/00 (2019.01 ) an automotive device , an identity and access management
( 58 ) Field of Classification Search (IAM) , or a robotic or high functioning touch sensing
CPC G06F 21/32 ; G06F 3/0487 ; G06K 9/6255 ; device .
GOON 20/00
See application file for complete search history. 16 Claims , 8 Drawing Sheets
100

104 106
GUI / Error
Username API Handling
-122 .114 116 -130
-120 Authentication
Data Security 108
Client / User's
Pressure Device Handling Property
Driver Server
Recorder -112
API Holder
Config / Config /
Control 110 Machine Data Control
Training Analysis
 US 11,281,755 B2
Page 2

( 56 ) References Cited [ online ] , [ Retrieved Nov. 13 , 2019 ] , Retrieved from the Internet
U.S. PATENT DOCUMENTS
< URL :https ://en.wikibooks.org/wiki/Robotics/Sensors/Pressure_
Sensors> (2 pgs) .
2016/0239201 A1 8/2016 Wang et al .
[ online ] , [Retrieved Nov. 13 , 2019 ] , Retrieved from the Internet
2016/0239652 A1 * 8/2016 Wang GOIL 5/0038 < URL :http : //www.lessemf.com/fabric.html#253 > ( 8 pgs ) .
2017/0243359 Al 8/2017 Bose et al . [ online ] , [ Retrieved Nov. 13 , 2019 ] . Retrieved from the Internet
2017/0322674 Al 11/2017 Rosenberg et al . < URL :http ://www.kobakant.at/DIY/? p = 213 > ( 9 pgs ).
2018/0263565 A1 * 9/2018 Najafi A61B 5/1123 International Search Report and Written Opinion dated Nov. 7 ,
2020/0380101 A1 12/2020 Arii 2019 , in related International Patent Application No. PCT/US/2019 /
047769 , filed Aug. 22 , 2019 ( 9 pgs ) .
OTHER PUBLICATIONS Non -Final Office Action dated Jan. 6 , 2021 in related U.S. Appl. No.
16 / 548,589 , (28 pgs ) .
[ online] , [Retrieved Nov. 13 , 2019 ] , Retrieved from the Internet Non - Final Office Action Response filed Mar. 4 , 2021 in related U.S.
Appl. No. 16/ 548,589 , ( 9 pgs ) .
< URL : https ://www.google.com/search ? q = bga + socket & tbm = isch Final Office Action dated Mar. 23 , 2021 in related U.S. Appl. No.
& source = iu & ictx = 1 & fir = bGxdm 2 KtOSSDfM % 253A % 16/ 548,589 , ( 24 pgs ) .
252C3b19ghq8UX2q8M % 252C_ & usg = Non -Final Office Action dated Sep. 14 , 2021 in related U.S. Appl.
AFrqEzeCA6efrw3Q4jsxJic70X0J_Pne4Q & s > , ( 16 pgs ) . No. 16/ 548,589 , ( 29 pgs ) .
[ online] , [Retrieved Nov. 13 , 2019 ] . Retrieved from the Internet
< URL :https : //www.peratech.com/m2436-100148-zf.html> , ( 5 pgs ) . * cited by examiner
U.S. Patent Mar. 22 , 2022 Sheet 1 of 8 US 11,281,755 B2

130
User's Prope ty Holder
Authenicao Client
/
C/onfig
Control
106 Server ???
116 112

100
Er or Handling 108
Analysi Data F1IGURE
/
GUI
API Security Handling Machine Train g
114
104 110
Device Driver
12 1
Config
/
120
D at a Control
Usernam Presure Recorde
U.S. Patent Mar. 22 , 2022 Sheet 2 of 8 US 11,281,755 B2
200

201
Input Buffer List
202
203 Input Buffer
Pixel Pressure value
Pixel Pressure value
202
Input Buffer

202
Input Buffer

FIGURE 2A
U.S. Patent Mar. 22 , 2022 Sheet 3 of 8 US 11,281,755 B2

250

204
Database Buffer List
205
206 Database Buffer
207 Database Buffer history values
Pixel Pressure history values :
Pixel Pressure history values
205
Database Buffer
205
:
Database Buffer

FIGURE 2B
U.S. Patent Mar. 22 , 2022 Sheet 4 of 8 US 11,281,755 B2

300

301
Hybrid Buffer List
302
306
303 Hybrid Buffer 307
304 Hot Spots Input buffer pointer *
308
305 Intensity Database buffer pointer *
Borders
302
Hybrid Buffer
302

Hybrid Buffer

FIGURE 3
U.S. Patent Mar. 22 , 2022 Sheet 5 of 8 US 11,281,755 B2

400

401
New Data Structure
402
302
403 New Buffer Group
404 Key buffer Hybrid Buffer
302
302 Priority
402
Hybrid Buffer Hybrid Buffer
New Buffer Group
402

New Buffer Group

FIGURE 4
U.S. Patent Mar. 22 , 2022 Sheet 6 of 8 US 11,281,755 B2

500

501
User Account
502
User Name

503

Properties need authentication

201
Database Buffer List
504
Valid compare value range

FIGURE 5
U.S. Patent Mar. 22 , 2022 Sheet 7 of 8 US 11,281,755 B2

LBHuyifbsretid
604 CDonaclutsiaon 614 NewHybrid dstrauctare -618
606
No
/
Yes
600 DPraoctesor DCoamptare 6FIGURE
610 616

DBLautifbsaetr LBIuinfspuetr
602
608 612
dNewInaptuat struc e
U.S. Patent Mar. 22 , 2022 Sheet 8 of 8 US 11,281,755 B2

DBLautifbsaetr
602
702 DNsatreucbwsre
700 DMaertgae 7FIGURE
704
610
PDroacetsaor
sdINtnrauepctuware
612
 US 11,281,755 B2
1 2
SYSTEMS AND METHODS FOR SECURE FIG . 3 illustrates an exemplary hybrid buffer list accord
BIOMETRIC IDENTIFICATION USING ing to various embodiments of the present disclosure .
RECORDED PRESSURE FIG . 4 illustrates an exemplary data structure according to
various embodiments of the present disclosure .
CROSS REFERENCE TO RELATED PATENT 5 FIG . 5 illustrates an exemplary user account according to
APPLICATIONS various embodiments of the present disclosure .
FIG . 6 is a flowchart of an illustrative process for secure
The present application claims priority benefit, under 35 biometric identification using recorded pressure according
U.S.C. $ 119 ( e ) , to commonly -assigned U.S. Patent Appli- 10
to various embodiments of the present disclosure.
cation No. 62 /730,948 , filed on Sep. 13 , 2018 , entitled FIG . 7 is aa flowchart of an illustrative process for merging
“ SYSTEMS AND METHODS FOR SECURE BIOMET- data according to various embodiments of the present dis
RIC IDENTIFICATION USING RECORDED PRES- closure.
SURE,” listing as inventors , Hong Chang and Alex Cheuk
Kwong Lee , which application is herein incorporated by 15
DETAILED DESCRIPTION OF EMBODIMENTS
reference as to its entire content. Each reference mentioned
in this patent document is incorporated by reference herein In the following description , for purposes of explanation,
in its entirety. specific details are set forth in order to provide an under
standing of the invention . It will be apparent, however, to
A. TECHNICAL FIELD 20 one skilled in the art that the invention can be practiced
without these details . Furthermore, one skilled in the art will
The present disclosure relates generally to systems and recognize that embodiments of the present invention,
methods for personal identification based on human -ma- described below, may be implemented in a variety of ways ,
chine interaction . More particularly, the present disclosure such as a process , an apparatus, a system , a device , or a
relates to biometric systems and methods that use dynamic 25 method on a tangible computer - readable medium .
pressure sensing to uniquely identify users . Components, or modules, shown in diagrams are illustra
tive of exemplary embodiments of the invention and are
B. BACKGROUND meant to avoid obscuring the invention . It shall also be
understood that throughout this discussion components may
Current biometric methods for personal identification use 30 be described as separate functional units, which may com
fingerprint data , retina scans, facial recognition methods, prise sub -units. Those skilled in the art will recognize that
and the like . All these identification methods have in com- various components, or portions thereof, may be divided
mon that they operate on static biometric data , i.e. , each of into separate components or may be integrated together,
the biological traits are unique for every individual and serve including integrated within a single system or component. It
as the basis for identification data that for practical purposes 35 should be noted that functions or operations discussed herein
does not change over time . One obvious drawback of static may be implemented as components . Components may be
biometric methods for personal identification is that aa moti- implemented in software, hardware, or a combination
vated attacker or identity thief can relatively easily circum thereof.
vent such methods by copying and then reproducing the Furthermore, connections between components or sys
identifying data to mimic the presence of the user having the 40 tems within the figures are not intended to be limited to
unique trait. For example , by presenting to a facial recog direct connections . Rather, data between these components
nition system a color copy of a picture of aa to -be - identified may be modified , re - formatted , or otherwise changed by
person , an imposter may successfully trick a security system intermediary components. Also , additional or fewer connec
into mistakenly recognizing the photo as that person , thus, tions may be used . It shall also be noted that the terms
defeating the security system and gaining access to confi- 45 " coupled , " " connected," or " communicatively coupled "
dential information . shall be understood to include direct connections, indirect
Accordingly, what is needed are dynamic systems and connections through one or more intermediary devices, and
methods that are virtually impossible to copy and, thus, wireless connections.
inherently more secure for safeguarding confidential infor- Reference in the specification to “ one embodiment,”
mation . 50 " preferred embodiment," " an embodiment," or " embodi
ments ” means that a particular feature , structure , character
BRIEF DESCRIPTION OF THE DRAWINGS istic , or function described in connection with the embodi
ment is included in at least one embodiment of the invention
References will be made to embodiments of the invention , and may be in more than one embodiment. Also , the
examples of which may be illustrated in the accompanying 55 appearances of the above -noted phrases in various places in
figures . These figures are intended to be illustrative , not the specification are not necessarily all referring to the same
limiting . Although the invention is generally described in the embodiment or embodiments .
context of these embodiments, it should be understood that The use of certain terms in various places in the specifi
it is not intended to limit the scope of the invention to these cation is for illustration and should not be construed as
particular embodiments. 60 limiting . A service , function , or resource is not limited to a
FIG . 1 is a simplified block diagram of a secure biometric single service , function, or resource ; usage of these terms
identification system according to various embodiments of may refer to a grouping of related services, functions, or
the present disclosure. resources, which may be distributed or aggregated. Further
FIG . 2A illustrates an exemplary input buffer list accord- more , the use of memory, database, information base , data
ing to various embodiments of the present disclosure . 65 store, tables , hardware, and the like may be used herein to
FIG . 2B illustrates an exemplary database buffer list refer to system component or components into which infor
according to various embodiments of the present disclosure. mation may be entered or otherwise recorded .
 US 11,281,755 B2
3 4
In this document “ sensor array” and “ electrical grid ” are and /or a characteristic common to two or more patterns. In
a
used interchangeably. Similarly, “ pressure" and " force " may embodiments, patterns may be selected from ( e.g. , random )
be used interchangeably, since as will be recognized by one samples and stored in a database , e.g. , in data analysis
of skilled in the art, pressure is defined as force applied onto module 112. The learned characteristics may then be used to
a given surface area. Therefore, the term “ pressure data” 5 identify the user that caused the pattern ( s) to be generated .
may refer to data that indicates pressure or a quantity, such In embodiments, once input data 122 has been received ,
as the physical quantity of force , from which pressure may machine learning may be applied to it in order to identify
be derived . The terms “ pressure sensor ” and “ pixel” refer to patterns and other characteristics therein to generate the
a device for directly or indirectly measuring , at one or more groups and sub -groups. In embodiments, each group and
locations, pressure or force, i.e. , a quantity from which 10 even each pixel may be assigned a weight factor, for
pressure data may be derived. example , based on a set of predetermined criteria . Machine
FIG . 1 is a simplified block diagram of a secure biometric training module 110 may adjust weight factors and other
identification system according to various embodiments of parameters over time , for example, to accommodate natural
the present disclosure . System 100 comprises GUI/ API 104 ,
error handling module 106 , security handling module 108 , 15 variations in a user's pressure and speed response. Machine
machine training module 110 , data analysis module 112 , training module 110 may apply and update training methods
device driver 114 , client/ server API 116 , and pressure at various times , for example, based on gestures or motion
2

recorder 120. It is understood that modules described herein that have been entered in to a database after being recog
may be implemented by a state machine that uses wired nized as valid .
logic or in software that executes instructions using a 20 Once the temporal and spatial characteristic of input data
processing unit such as a microcontroller. 122 is successfully mapped to aa known gesture in a database ,
An example of pressure sensors and pressure recorder 120 a user may be recognized and the identification process may
that may benefit from the teachings of the present disclosure conclude . In embodiments, machine training module 110
are the systems and methods disclosed in U.S. Provisional accounts for tolerances, such as pressure variations caused
Patent Application , Ser. No. 62 /729,738 , entitled “ Pressure 25 by temperature and other environmental parameters that,
Recording Systems and Methods for Biometric Identifica- otherwise , may negatively impact sensor accuracy and , thus,
tion ,” filed on Sep. 11 , 2018 , 2018 and listing as inventors the accuracy of pressure recorder 120. It is understood that
Alex Cheuk Kwong Lee and Hong Chang, which patent tolerances may also be user - related , e.g. , due to factors such
application is incorporated herein by reference as to its as the user's physical condition, which may vary over a
entire content and for all purposes. 30 length of time whether within a day or within a decade .
In embodiments, input data 122 may comprise data pro- The learning of patterns and characteristics may be per
vided by a multi -dimensional array of pressure sensors , e.g. , formed by machine training module 110 that receives input
spatial and temporal pressure data, such that the data cap- data 122 as training data . In embodiments, in a training
tured by pressure recorder 120 may represent a spatial phase, for example , after in initialization procedure that may
pressure distribution that may vary in time . In embodiments, 35 involve prompting the user to provide , via a user interface
input data 122 may be based on fusing data fromany number 104 ( not shown) , a username, the user may be prompted to
of different types of sensors . Temporal variations in pressure exert pressure on the pressure sensor using his thumb , finger,
data may be continuous or, e.g. , according to discrete or any other body part.
periodic time intervals . In embodiments, in order to ensure In embodiments, the user may be prompted to repeatedly
accuracy , input data 122 may comprise pressure measure- 40 press the sensor a number of times , e.g. , with varying force,
ment data saved in a time sequence as pixel buffer values . speed , and position , e.g. , by rotating or rolling the thumb ,
It is noted that pressure recorder 120 may be a high- resulting in different angles and/ or sequences at which the
resolution pressure recorder 120 that may accurately cap- pressure sensor captures user -generated pressure -related
ture, e.g. , at dynamically changing measurement rates , the data, such as temporal and spatial information , and envi
nuances of even unconsciously exerted gestures or move- 45 ronmental parameters, such as temperature, etc. , that may
ments when a user interacts with pressure sensor 120 , e.g. , then be pre - processed and stored in a database , e.g. , within
using a finger or another body part. In embodiments, data pressure recorder 120 .
analysis module 112 may apply relational analysis to infer or In embodiments, machine training module 110 may use
derive data 122 from a recorded movement by the user of averaging techniques, for example , to average pressure data
pressure recorder 120 , such as the user's rolling of a thumb 50 122 over a period of time , area , or both , e.g. , to determine
while exerting pressure on a pressure touch sensor. a mean location of a pressure distribution ( similar to a center
In embodiments, data 122 may be used to generate a of gravity ) in each frame or region of a frame. Then an
recording ( e.g. , frame -by - frame) of pressure data that rep- activity level may be defined and assigned to one or more
resents a spatial distribution , such that each frame may be regions within each frame. For example, pressure data
assigned a number of ( e.g. , different) patterns or sub - patterns 55 within a region may be averaged to determine the highest
associated with a number of pressure regions. In embodi- number or active pixels , i.e. , pixels having values above a
ments , input data 122 may be pre -processed, e.g. , sorted , certain threshold . In embodiments, within a given frame or
grouped, and processed in a manner that facilitates the based on data continuity, pixels may be defined as active
comparison of patterns to each other. In embodiments , based on a number of less active neighboring pixels . In
pre -processing comprises fusing data from two or more 60 embodiments, a number of active pixels may be clustered
sources that may comprise additional pressure - related data . into sub -groups or regions that may be collectively catego
Similarly, data from aa number of regions may be processed rized as an “ active region . ” In embodiments , regions may be
before being used as training data and stored in the database . combined to form a data structure that represents a user
In embodiments , input data 122 may be used to learn a set profile that may then be used to identify the user. In
of ( sub- )patterns that may be associated, e.g. , with physical 65 embodiments, a set of most active pixels may be used to map
parameters, such speed , range , etc., and that may be used to one orientation into another to account for tilting of a
identify a number of characteristics, e.g. , within the pattern reference frame.
 US 11,281,755 B2
5 6
In embodiments, a database may store historical data that 2B illustrates an exemplary database buffer list according to
may serve as contextual data for interpreting or evaluating various embodiments of the present disclosure . FIG . 3
the significance of sensor data from pressure recorder 120 . illustrates an exemplary hybrid buffer list according to
For example, having knowledge of the fact that the user is various embodiments of the present disclosure. FIG . 4
a child or elderly person when generating a user profile may 5 illustrates an exemplary data structure according to various
aid in more accurately interpreting data 122 , for example, by embodiments of the present disclosure . FIG . 5 illustrates an
taking into account that certain categories of users generally exemplary user account according to various embodiments
exert less pressure on the sensor when so prompted. Simi- of the present disclosure .
larly, a child with not fully developed motor skills may FIG . 6 is a flowchart of an illustrative process for secure
behave less consistent and, thus, generate responses having 10 biometric identification using recorded pressure according
more variation when asked to press the sensor five times in to various embodiments of the present disclosure. In
a row compared to an adult. embodiments, data conclusion module 604 receives histori
Therefore , in embodiments, historical data may be used as cal pixel pressure values from database buffer list 204 or
a factor in determining a confidence interval that may define pixel values from input buffer list 201 and computes there
what would be considered reliable data. Conversely, in 15 from a relative relation of pixels to their corresponding
embodiments, based on the assumption that no person can neighbors. It is understood that the computations may be
reproduce two identical gestures by applying pressure to a performed for some or all pixel in a buffer. In embodiments,
surface, receiving identical or highly accurate data may be one or more threshold values are used to filter or sort pixel
used as an indicator that an attacker uses somehow dupli- data 207 or 203 , e.g. , to find pixel boundaries 305 , hot spots
cated or copied real - time pressure data to simulate aa certain 20 303 , intensity data 304 , or a VIX data 308 for the buffer. Any
user's individual pressure characteristic . Thus , in embodi- of the filtered /sorted data may be combined into a hybrid
ments, the confidence interval may be defined by a volatility package 302. In embodiments, the steps beginning with the
>
index ( VIX ), i.e. , range of expected accuracy or values that computation may be repeated for each buffer 205 or 202 to
may account for sample - to - sample variations. Machine output a list 301 of hybrid buffers 302 that may comprise
training module 110 may calculate the VIX , e.g. , in the 25 information about pixel boundaries 305 , hot spots 303 ,
training phase , based on differences in the user input data , as intensity data 304 , and VIX data 308 for each hybrid buffer
a user -specific confidence range. In embodiments, to further 302. In embodiments, hybrid buffer 302 may comprise input
increase security, security handling module 108 may , among buffer pointers 306 and database buffer pointers 307 that
other things , handle user access rights using any security may compose a null pointer.
model known in the art, and adjust confidence levels based 30 In embodiments, data processor 610 receives input from
on expected environments and user - related tolerances . at least one of input buffer list 201 , hybrid buffer list 301 , or
In embodiments, a time-sequenced compare database 602 database buffer list 204. When the input is not received from
may be used to generate a more compact hybrid compare hybrid buffer list 301 , then it may be received from input
database 606 that may than be communicated to a client side buffer list 201 or database buffer list 204 .
server where it may be reorganized into one or more 35 In embodiments , data processor uses list 301 in data
non- time based groups based on Al-defined characteristics conclusion module 604 to find a key buffer having the
614. In addition, the format of the database may be chosen largest border, the most hot spots , or the most intensity
to be significantly different from input data 122 , such that among hybrid buffers 302. In embodiments, data processor
there is no way to mimic input data 122 to circumvent the 610 reconstructs the buffers 302 into new groups 402 and
security of system 100. In embodiments, each data's char- 40 assigns priorities 404 the groups , e.g. , based on a calculated
acter may be based on neighboring data and the data value representative of a defined characteristic . In embodi
collected immediately prior or after. In embodiments , small ments, groups 402 may be divided into sub - groups that each
random adjustments in input data 122 may significantly undergo a similar process in a recursive function until new
change the data character and , thus, trigger an identification data structure 401 is output. New data structure 401 may or
failure. 45 may not have aa handle point back to original buffers 205 or
In embodiments, input data 122 changes the content of the 202 input to the function .
database buffer list 602 , hybrid buffer list 606 , and the new In embodiments, data compare module 616 receives as
hybrid data structure 614 , such that if an attacker manages input at least one of hybrid buffer list 301 , new input hybrid
to steal information contained in these databases, the static buffer list 301 with a handle pointed back the input buffer
information would become obsolete . In embodiments, 50 202 ; or a valid compare value range 504 for a specific user .
authentication does not depend upon a straightforward com- In embodiments, data compare module 616 aligns key buffer
parison of these databases, further thwarting any attackers . 403 , compares each group 402 , compares result multiply
In embodiments, to reduce computation time , only partial priority 404 , and each group 402 result together, and com
necessary database may be communicated to the client side. pares the calculation result to range 504 for the user . Upon
This keeps memory allocation at the client side reasonably 55 the calculation values falling within the range process 600
small and reduces bandwidth requirements when transfer- may resume with entering a data merge process described
ring data over the internet. In embodiments, to reduce a next.
computational burden , a number of frames that have the FIG . 7 is aa flowchart of an illustrative process for merging
highest number of active pixels may be used for evaluation data according to various embodiments of the present dis
purposes. In embodiments, to accomplish this , pressure data 60 closure. In embodiments, data merge procedure 700 com
within aa frame may be summed to obtain the frame having prises receiving input data from new data structure 401
the most active points, i.e. , the frame associated with the having a handle pointed back to database buffer 205 and new
highest total pressure. Then the frame may be used as the data structure 401 with aa handle pointed back to input buffer
basis for generating a unique profile /highest resolution char- 202 .
acteristic . 65 In embodiments, data merge process 700 aligns key buffer
FIG . 2A illustrates an exemplary input buffer list accord- 403 , and separately merges each new buffer group 402. In
ing to various embodiments of the present disclosure . FIG . embodiments, if two new data structures 401 do not have the
 US 11,281,755 B2
7 8
same length within group 402 , the respective buffer 205 may intended that all permutations, enhancements, equivalents,
be trimming or padded , and a decision may be made by combinations, and improvements thereto that are apparent to
which one have stronger character. those skilled in the art upon a reading of the specification
In embodiments, procedure 700 comprises mapping in and a study of the drawings are included within the true
each frame each pixel 207 to historic pressure values 203 5 spirit and scope of the present disclosure . It shall also be
and recalculating pixel history values 207 , recalculating noted that elements of any claims may be arranged differ
database buffer history values 206 , and / or recalculating ently including having multiple dependencies, configura
valid compare value range 504 for a given user. In embodi- tions , and combinations.
ments, new buffer groups 402 may be added together to form What is claimed is :
an updated new database buffer list 204. In embodiments , 10 1. A biometric identification method comprising:
the results may be written into database buffer list 204 . receiving input data comprising spatial and temporal
Aspects of the present invention may be encoded upon pressure data that has been generated in response to a
one or more non - transitory computer - readable media with user exerting pressure on a pressure sensor by pressing
instructions for one or more processors or processing units down the pressure sensor, the pressure sensor compris
to cause steps to be performed. It shall be noted that the one 15 ing a plurality of regions, wherein the at least some of
or more non -transitory computer - readable media shall the spatial and temporal pressure data has been
include volatile and non - volatile memory . It shall be noted recorded in frames, each frame being assigned an
that alternative implementations are possible , including a activity level based on a set of active pixels , the activity
hardware implementation or a software / hardware imple level being defined by pressure data that has been
mentation . Hardware - implemented functions may be real- 20 averaged within a region to determine the highest
ized using application specific integrated circuits ( ASICs ) , number of pixels exceeding a threshold ;
programmable arrays, digital signal processing circuitry, or using the set of active pixels to map an orientation into
the like . Accordingly , the terms in any claims are intended another orientation to account for tilting of a reference
to cover both software and hardware implementations. Simi- frame;
larly, the term “ computer -readable medium or media ” as 25 using the input data to identify one or more patterns
used herein includes software and /or hardware having a associated with the plurality of regions;
program of instructions embodied thereon, or a combination associating the one or more patterns with one or more
thereof. With these implementation alternatives in mind, it is characteristics; and
to be understood that the figures and accompanying descrip- using the one or more characteristics to identify the user .
tion provide the functional information one skilled in the art 30 2. The biometric identification method according to claim
would require to write program code ( i.e. , software) and / or 1 , further comprising:
to fabricate circuits ( i.e. , hardware) to perform the process- comparing the one or more patterns to patterns that have
ing required . been obtained in a training phase by a trained model
It shall be noted that embodiments of the present inven that is updated based on successfully learned charac
tion may further relate to computer products with a non- 35 teristics;
transitory , tangible computer - readable medium that have based on the comparison, calculating a probability that the
computer code thereon for performing various computer one or more patterns match the one or more character
implemented operations. The media and computer code may istics; and
be those specially designed and constructed for the purposes in response to the probability satisfying a threshold ,
of the present invention , or they may be of the kind known 40 authenticating the user .
or available to those having skill in the relevant arts. 3. The biometric identification method according to claim
Examples of tangible computer - readable media include, but 1 , wherein the input data comprises fused data from two or
2

are not limited to : magnetic media such as hard disks; optical more sensors .
media such as CD - ROMs and holographic devices ; mag- 4. The biometric identification method according to claim
neto -optical media ; and hardware devices that are specially 45 1 , wherein the spatial and temporal pressure data has been
configured to store or to store and execute program code, acquired at varying sampling rates .
such as ASICs , programmable logic devices ( PLDs ) , flash 5. The biometric identification method according to claim
memory devices, and ROM and RAM devices. Examples of 1 , further comprising clustering pixels the set of active
computer code include machine code, such as produced by pixels into an active region.
a compiler, and files containing higher level code that are 50 6. The method according to claim 1 , further comprising
executed by a computer using an interpreter. Embodiments assigning to elements in the plurality of regions weight
of the present invention may be implemented in whole or in factors based on a set of predetermined criteria .
part as machine - executable instructions that may be in 7. The biometric identification method according to claim
program modules that are executed by a processing device . 6 , further comprising adjusting the weight factors to account
Examples of program modules include libraries, programs, 55 for at least one of a user - related or not user - related tolerance
routines, objects, components , and data structures . In dis- that causes variations in the pressure.
tributed computing environments, program modules may be 8. The biometric identification method according to claim
physically located in settings that are local , remote , or both . 7 , further comprising, in response to receiving data that is
One skilled in the art will recognize no computing system identical or highly accurate, using the data as in indication
or programming language is critical to the practice of the 60 of an attack .
present invention . One skilled in the art will also recognize 9. The biometric identification method according to claim
that a number of the elements described above may be 1 , wherein using the input data comprises using contextual
physically and /or functionally separated into sub -modules or data to evaluate the input data .
combined together. 10. The biometric identification method according to
It will be appreciated to those skilled in the art that the 65 claim 1 , further comprising evaluating sample - to -sample
preceding examples and embodiments are exemplary and variations by comparing samples with a range of expected
not limiting to the scope of the present disclosure . It is data .
 US 11,281,755 B2
9 10
11. The biometric identification method according to associating the one or more patterns with one or more
claim 1 , further comprising selecting from among the spatial characteristics; and
and temporal pressure data a set of data to serve as an using the one or more characteristics to identify the user.
orientation reference data set . 13. The biometric identification system according to
12. A biometric identification system comprising: 5 claim 12 , wherein the biometric identification system is used
a pressure recorder that comprises a pressure sensor that to authenticate or identify the user in connection with at least
comprises a plurality of regions, the pressure sensor one of an electronic commercial transaction , an Internet of
receives input data comprising spatial and temporal accessThings (IoT ) device , an automotive device , an identity and
management (IAM) , or a robotic or high functioning
pressure data that has been generated in response to a 10 touch sensing
user exerting pressure on the pressure sensor by press device .
14. The biometric identification system according to
ing down the pressure sensor, the pressure event being claim 12 , further comprising
recorded and used to generate a user profile, wherein rality of regions weight factorsassigning
based
to elements in plu
on a set of predeter
the spatial and temporal pressure data has been
recorded in frames, each frame being assigned an mined criteria, and adjusting the weight factors to account
activity level based on a set of active pixels in an active 15 for at least one of a user-related or notuser-related tolerance
that causes variations in the pressure.
a region, the activity level being defined by pressure 15. The biometric identification system according to
data that has been averaged within the region to deter claim 12 , further comprising reducing a computational bur
mine the highest number of pixels exceeding a thresh den by selecting among the frames those frames that have
old ;
using the set of active pixels to adjust a spatial orientation 20 the
old .
highest number of active pixels that exceed the thresh
to account for tilting of a reference frame; and 16. The biometric identification system according to
one or more processors that, in response to receiving the claim 12 , further comprising filtering pixel data to determine
spatial and temporal pressure data , performs steps
comprising: pixel boundaries for aa buffer that is used to perform one or
using a trained model that is updated based on success 25 prising
more computations to output a list of hybrid buffers com
information about the pixel boundaries .
fully learned characteristics to identify one or more
patterns associated with the plurality of regions ;