DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Preliminary Amendment
This Office Action is responsive to the amendment filed on 18 Jun 2026. As directed by the amendment: claims 1-2, 7-9, and 14 have been amended, claims 3-5 and 10-12 have been canceled, and claims 15-18 have been added. Thus, claims 1-2, 6-9, and 13-18 are presently pending in this application.
Response to Arguments
Claim Objections
Applicant’s arguments, see Remarks, filed 18 Jun 2026, with respect to the objections to the claims have been fully considered and are persuasive in light of the claim amendments. The objections to the claims have been withdrawn.
Claim Rejections - 35 USC § 101
Applicant's arguments filed 18 Jun 2026 have been fully considered but they are not persuasive.
Regarding Step 2A, Prong One, Applicant argues:
Claim 1 as amended now requires "applying at least one linear mapping comprising a plurality of bandpass filters" to temporal segments and "applying at least one nonlinear mapping comprising at least one of: rectification; exponentiation; and gain control" to produce a heartbeat frequency encoding. Applying a plurality of bandpass filters combined with nonlinear mapping operations (rectification, exponentiation, gain control) to radar-derived physiological signal segments cannot practically be performed mentally or with pen and paper. Furthermore, claim 1 as amended requires applying "a trained machine learning subject classification model". A trained model is by definition a computational model that has been trained on data, involving trained parameters and algorithmic decision-making, not a mental comparison.
(Remarks, page 9)
Examiner respectfully disagrees. “applying a plurality of bandpass filters” is a mathematical operation. “Nonlinear mapping operations” comprising rectification, exponentiation, or gain control are also mathematical operations. Therefore, these are processes that can be performed mentally or with pen and paper. The algorithms used by a trained model are, by definition, mathematical operations, and are therefore also processes that can be performed mentally or with pen and paper.
Regarding Step 2A, Prong Two, Applicant argues:
Here, the claimed combination provides a technological solution to the technological problem of contactless biometric identification. This is analogous to Thales Visionix Inc. v. United States, 850 F.3d 1343, 1348-49 (Fed. Cir. 2017), where the particular configuration of inertial sensors and the particular method of using raw data from the sensors was found to be more than simply applying a law of nature because the claims provided a system that eliminated complications inherent in previous solutions. Similarly, claim 1 as amended recites a particular configuration of a THz radar detector and a particular method of processing the raw radar data (i.e., deriving BCG signals, segmenting into individual heartbeat segments, applying both linear and nonlinear mappings) to produce heartbeat frequency encodings for trained ML-based biometric identification, providing a specific technological solution rather than merely applying an abstract idea on a generic computer.
(Remarks, page 10)
Examiner respectfully disagrees. The claims at issue in Thales were focused on specific systems and methods that use inertial sensors in a non-conventional manner to reduce errors in measuring the relative position and orientation of a moving object on a moving reference frame. According to the court, the Thales claims specify a particular configuration of inertial sensors and a particular method of using the raw data from the sensors in order to more accurately calculate the position and orientation of an object on a moving platform. The mathematical equations are a consequence of the arrangement of the sensors and the unconventional choice of reference frame in order to calculate position and orientation. Far from claiming the equations themselves, the claims seek to protect only the application of physics to the unconventional configuration of sensors as disclosed.
However, none of claim 1 is directed to an unconventional configuration of sensors or a specific process for reducing errors in measuring the relative position and orientation of a moving object on a moving reference frame as discussed in Thales. Thus, the applicability of Thales to these claims is questionable, at best. Further, unlike the claims at issue in Thales, claim 1 merely applies an abstract idea to a computer and do not either improve the performance of the computer itself or computer technology in any way. Therefore, claim 1 recites an abstract idea.
Applicant argues:
Claim 1 as amended recites "in response to classifying the heartbeat frequency encoding as belonging to the reference subject, performing at least one action selected from: associating collected cardiac signal data of the monitored subject with stored information of the reference subject; or generating a notification based on the identification of the monitored subject". This is analogous to USPTO Subject Matter Eligibility Example 47, Claim 3, which was found eligible because the claim included specific actions taken based on the output of the ML model ( dropping malicious packets, blocking future traffic) that integrated the abstract idea into a practical application of improving network security (See Example 47, Claim 3: "Steps (d)-(f) provide for improved network security using the information from the detection to enhance security by taking proactive measures to remediate the danger...Thus, the claim as a whole integrates the judicial exception into a practical application such that the claim is not directed to the judicial exception"). Similarly, claim 1 as amended takes concrete action based on the classification result, namely: associating cardiac data with stored information or generating a notification, which integrates any alleged abstract idea into the practical application of biometric identification. These post-classification actions cannot be performed mentally (See MPEP 2106.04(d)(l) and 2106.05(a) (improvements to technology); MPEP 2106.05(e) (other meaningful limitations)).
(Remarks, pages 10-11)
“Associating cardiac data with stored information” is a mental process of linking the cardiac data with previously gathered information. “Generating a notification” encompasses nothing more than displaying a result of data analysis, which is not sufficient to show an improvement to computing technology, as per MPEP 2106.05(a)(II).
Applicant argues that “Islam et al. does not disclose the specific combination of THz radar with BCG-derived heartbeat frequency encodings produced through both linear and nonlinear mappings, trained ML classification, and concrete post classification actions as now recited in the amended claims” (Remarks, page 11). Islam was referenced in relation to known methods of remotely measuring vital signs using radar, not as a prior art rejection of each of the data processing steps of claim 1.
Therefore, claims 1-8 remain rejected under 35 U.S.C. 101 below.
Claim Rejections - 35 USC§ 103
Applicant’s arguments, see Remarks, filed 18 Jun 2026 , with respect to the rejection(s) of claim(s) 1 and 8 under 35 U.S.C. 103 have been fully considered and are persuasive in light of the claim amendments. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Frady et al. (Computing on Functions Using Randomized Vector Representations, 08 Sep 2021), hereinafter Frady, as explained in further detail below.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-2, 6-9, and 13-18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more.
Determination as to whether a claim satisfies the criteria for subject matter eligibility is a stepwise process (MPEP 2016).
Step 1: Does the claim fall within a statutory category of invention?
Claim 1 recites a process (method), and claim 8 recites a machine (system), which are within the four statutory categories. Therefore, claims 1 and 8 are directed to a statutory category of invention.
Step 2A, Prong 1: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Claims 1 and 8 are directed to an abstract idea.
Claim 1 is directed to receiving a reflection radar signal reflected from a body tissue of a monitored subject using a remote non-invasive radar detector, wherein the reflection radar signal is in the frequency range of 0.03 to 3 THz; deriving a cardiac ballistocardiogram (BCG) signal from the reflection radar signal; segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat; applying at least one linear mapping comprising a plurality of bandpass filters to each of the temporal segments, and applying a trained machine learning subject classification model to the heartbeat frequency encoding to classify the heartbeat frequency encoding as belonging to a reference subject of a reference subject group if a matching classification is obtained or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no machine classification is obtained; and in response to classifying the heartbeat frequency encoding as belonging to the reference subject, performing at least one action selected from: associating collected cardiac signal data of the monitored subject with stored information of the reference subject; or generating a notification based on the identification of the monitored subject.
Claim 8 recites a system that carries out the same method disclosed in claim 1.
The limitations of segmenting the derived BCG signal, applying at least one linear mapping, applying at least one nonlinear mapping, and applying at least one machine learning model, as drafted, under their broadest reasonable interpretations, are merely mental processes, because these steps are akin to having a doctor or other human actor performing these operations with pen and paper. For example, “segmenting the derived BCG signal” encompasses nothing more than a human actor mentally evaluating the collected data and deciding how to segment it. The limitation of “applying at least one machine learning model…to classify the heart beat frequency encoding” encompasses nothing more than a human actor mentally evaluating the heartbeat frequency encoding by comparing it to a reference to reach a conclusion about the encoding.
Therefore, claims 1 and 8 recite an abstract idea.
Claims 2, 6-7, 15, and 17 depend on claim 1, and claims 9, 13-14, 16, and 18 depend on claim 8. These dependent claims only recite additional features of the analysis described in claims 1 and 8, which may also be performed by a human actor mentally and using a pen and paper. For example, claims 7 and 14 recite “simultaneously monitoring or identifying multiple subjects in a location using source separation techniques”, which encompasses nothing more than a human actor mentally evaluating source separation techniques in order to determine the separate sources.
Therefore, claims 1-2, 6-9, and 13-18 recite an abstract idea.
Step 2A, Prong 2: Does the claim recite additional elements that integrate the judicial exception into a practical application?
This judicial exception is not integrated into a practical application.
Claim 8 only recites the additional limitations “a cardiac signal processor” and “a machine learning processor”. These additional elements are recited at a high level of generality (i.e. most generic computers would be known to have these components). Pages 12-13 of the specification describe the processors at a high level of generality. These generic processor limitations are no more than mere instructions to apply the exception using a generic computer component. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. Therefore claim 10 does not integrate the judicial exception into a practical application.
Claim 8 recites the additional limitation “a cardiac signal detector”, which amounts to no more than mere pre-solution activity of data gathering. Therefore the claimed generic detector element does not integrate the judicial exception into a practical application.
Claims 5 and 12 recite the additional limitation of “a remote non-invasive radar detector”, which amounts to no more than pre-solution activity of data gathering. Therefore, the claimed generic detector does not integrate the judicial exception into a practical application.
Claims 1 and 8 recite “in response to classifying the heartbeat frequency encoding as belonging to the reference subject, performing at least one action selected from: … generating a notification based on the identification of the monitored subject”, which encompasses nothing more than displaying a result of data analysis, which is not sufficient to show an improvement to computing technology, as per MPEP 2106.05(a)(II).
Thus, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Therefore, the claims are directed to an abstract idea. As described above, dependent claims 2-4, 6-7, 9-11, and 13-14 only recite other limitations of processing and analyzing the cardiac signal, which may be done mentally by a human actor and/or with a pen and paper.
Step 2B: Does the claim include additional elements that are sufficient to amount to significantly more than the judicial exception?
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As explained above with respect to the integration of the judicial exception into a practical application (Step 2A, Prong 2), the additional elements of using computer components to perform the process steps amounts to no more than mere instructions to apply the judicial exception using generic computer elements. The structural elements recited in claim 8 are “a cardiac signal processor”, and “a machine learning processor”. These additional elements are recited at a high level of generality (i.e. most generic computers would be known to have these components). Pages 12-13 of the specification describe the processors at a high level of generality, and only provides conventional, well-known computing functions that do not add meaningful limits to practicing the abstract idea.
Claims 5 and 12 recite the additional limitation “a remote non-invasive radar detector”. As discussed above with respect to integration of the abstract idea into a practical application (Step 2A, Prong 2), the additional element of a detector to collect data amounts to no more than mere pre-solution activity of data gathering. This pre-solution activity of data gathering using a remote non-invasive radar detector is well-understood, routine, and conventional in the field of radar-based identity authentication technology. For example, see Islam et al. (“Radar-Based Non-Contact Continuous Identity Authentication”, 2020, cited in IDS filed 02 Jan 2026), which describes known methods of remotely measuring vital signs using radar (Section 3.2. Radar-Based Identity Authentication through Heart-Based Features). Therefore, the claimed generic radar detector and computer processing elements are all well-understood, routine, and conventional in the field of radar-based identity authentication technology.
Therefore, claims 1-2, 6-9, and 13-18 are not patent-eligible under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-2, 6-9, and 13-18 are rejected under 35 U.S.C. 103 as being unpatentable over Steinberg et al. (US 20220079464 A1, previously cited), hereinafter Steinberg, in view of Liu et al. (US 20170188971 A1, previously cited), hereinafter Liu, and further in view of Frady et al. (Computing on Functions Using Randomized Vector Representations, 08 Sep 2021), hereinafter Frady.
Regarding claim 1, Steinberg discloses a method for biometric identification (paragraphs [0100], [0108]), the method comprising the procedures of:
receiving a reflection THz radar signal reflected from a body tissue of a monitored subject using a remote non-invasive radar detector, wherein the reflection radar signal is in the frequency range of 0.03 to 3 THz (paragraph [0080], "a remote portable non-contact detection system ... one or more reception means for receiving the sub-THz and THz signal of the subject. The received sub-THz and THz signals being a reflection of the sub-THz and THz signal from subject thereby"; paragraph [0088] ,"electromagnetic waves within the ITU-designated band of frequencies from 0.03 to 3 terahertz");
deriving a cardiac ballistocardiogram (BCG) signal from the reflection THz radar signal (paragraph [0090], "The microprocessor is further configured to perform analysis, calculation, data processing, automated reasoning, storing and/or processing the received sub-THz and THz signals and detect at least one physiological parameter"; paragraph [0088], "The term ‘Physiological Parameters’ herein refers to any physiological indicator...such as...ballistocardiogram(BCG), BCG amplitude variability"); and
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment being of a selected time duration (paragraph [0133], "folding or mirroring the signals and decimating selected portions of the folded signals and removing folded segments");
applying at least one linear mapping comprising a plurality of bandpass filters to each of the temporal segments (paragraph [0133], bandpass filtering);
identifying a subject based on the BCG signal according to stored data (paragraphs [0100], [0108], [0113]); and
in response to classifying the heartbeat frequency encoding as belonging to the reference subject, performing at least one action selected from: associating collected cardiac signal data of the monitored subject with stored information of the reference subject (paragraph [0091], "The microprocessor may further detect, compare and provide interpretation of the received signal indicating of a change in subject's health condition, based on the received signal information and/or stored information"; paragraphs [0019], [0109]); or generating a notification based on the identification of the monitored subject (paragraph [0107], "the system may further export an output associated with the health status selected from the group consisting of an alert, an indication flag, an activation instruction of electronic device, electronic message or any combination thereof"; paragraphs [0015], [0021], [0042], [0051], [0124]).
Steinberg does not explicitly disclose segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat; applying at least one nonlinear mapping comprising at least one of: rectification; exponentiation; and gain control, to the temporal segments to produce a heartbeat frequency encoding; nor applying a trained machine learning subject classification model to the heartbeat frequency encoding to classify the heartbeat frequency encoding as belonging to a reference subject of a reference subject group if a matching classification is obtained or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no machine classification is obtained.
However, Liu teaches a method for ECG authentication (Abstract) comprising:
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat (Fig. 2, paragraph [0055], step 220, "The preprocessing includes detection of a fiducial point and acquirement of a data segment"; paragraph [0067], acquirement of data segments; paragraph [0100], "The data segments acquired based on the point R includes overall information associated with a single heartbeat");
applying at least one linear mapping comprising a plurality of bandpass filters to each of the temporal segments (Fig. 2, paragraph [0055], step 220, "the ECG authentication apparatus filters the ECG signal using a band pass filter configured to pass a predefined frequency band"; paragraphs [0017]-[0018], [0065], [0095]-[0097]);
applying at least one nonlinear mapping to the temporal segments to produce a heartbeat frequency encoding (paragraph [0073], "The deep training method indicates a machine learning algorithm for attempting a high-level abstraction by combining various non-linear transformation schemes"; paragraph [0080], "The identification signal is used for determining whether a result of identification performed on an entity of a type corresponding to the ECG training data through a nonlinear mapping of layers is valid. The verification signal is used for determining whether a result obtained by verifying whether two items of ECG training data belong to the same entity through the nonlinear mapping of the layers is valid");
applying at least one machine learning model for subject classification on the heartbeat frequency encoding during an identification stage to classify the heartbeat frequency encoding as belonging to a reference subject if a matching classification is obtained, or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no matching classification is obtained (Fig. 2, paragraph [0061], step 240, "The ECG authentication apparatus calculates a similarity between the semantic feature and a predefined registered feature or a reference feature corresponding to a target to be compared with the semantic feature. The authentication apparatus determines an authentication result to be a success in authentication or a fail in authentication based on a comparison result of the calculated similarity and a threshold").
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg with the teachings of Liu so that the method comprises applying at least one nonlinear mapping, to the temporal segments to produce a heartbeat frequency encoding; nor applying a trained machine learning subject classification model to the heartbeat frequency encoding to classify the heartbeat frequency encoding as belonging to a reference subject of a reference subject group if a matching classification is obtained or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no machine classification is obtained, because doing so provides a method of authentication using markers that are not easily stolen, lost, or forged (Liu, paragraph [0005]).
Liu does not explicitly disclose that the at least one nonlinear mapping comprises at least one of: rectification; exponentiation; and gain control.
However, Frady teaches methods of encoding and computing (Abstract) comprising applying at least one nonlinear mapping comprising exponentiation (page 2, "The function domain is encoded by exponentiating a fixed random base vector"; page 15, exponential map).
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg and Liu with the teachings of Frady so that the at least one nonlinear mapping comprises exponentiation, because doing so provides transparency and insight into the machine learning model's features and weights, which helps explain the decisions of a neural network in specific application settings (Frady, page 1, Introduction, first paragraph; page 2, "VFA is fully transparent. Vectors can represent individual data points as well as elements of a function space that is well-defined as a reproducing kernel Hilbert space. The domain of the functions can encode continuous-valued quantities in data, such as position, time or wavelength"; page 21, "Changing the input encoding from random projection methods in conventional reservoir computing to KLPEs will yield a new class of recurrent networks for computing in a transparent fashion in a well-defined function space").
Regarding claim 2, the method of claim 1 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses:
for each of a plurality of reference subjects (paragraphs [0109], [0115]),
receiving a reflection THz radar signal reflected from a body tissue of a monitored subject using contactless detection (paragraph [0080], "a remote portable non-contact detection system ... one or more reception means for receiving the sub-THz and THz signal of the subject. The received sub-THz and THz signals being a reflection of the sub-THz and THz signal from subject thereby");
deriving a cardiac ballistocardiogram (BCG) signal from the reflection THz radar signal (paragraph [0090], "The microprocessor is further configured to perform analysis, calculation, data processing, automated reasoning, storing and/or processing the received sub-THz and THz signals and detect at least one physiological parameter"; paragraph [0088], "The term ‘Physiological Parameters’ herein refers to any physiological indicator...such as...ballistocardiogram(BCG), BCG amplitude variability"); and
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment being of a selected time duration (paragraph [0133], "folding or mirroring the signals and decimating selected portions of the folded signals and removing folded segments");
applying at least one linear mapping to each of the temporal segments to produce a heartbeat frequency encoding (paragraph [0133], bandpass filtering; page 19, line 15 of the instant specification discloses that linear mapping can include band pass filtering); and
assigning the heartbeat frequency encoding to an identification label relating to the reference subject (paragraphs [0100], [0108], [0113]).
Steinberg does not explicitly disclose applying at least one nonlinear mapping comprising at least one of: rectification; exponentiation; and gain control, to the temporal segments to produce a heartbeat frequency encoding; nor forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects; and applying at least one machine learning process to the training dataset, to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification.
However, Liu further teaches:
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat (Fig. 2, paragraph [0055], step 220, "The preprocessing includes detection of a fiducial point and acquirement of a data segment"; paragraph [0067], acquirement of data segments; paragraph [0100], "The data segments acquired based on the point R includes overall information associated with a single heartbeat");
applying at least one nonlinear mapping to the temporal segments to produce a heartbeat frequency encoding (paragraph [0073], "The deep training method indicates a machine learning algorithm for attempting a high-level abstraction by combining various non-linear transformation schemes"; paragraph [0080], "The identification signal is used for determining whether a result of identification performed on an entity of a type corresponding to the ECG training data through a nonlinear mapping of layers is valid. The verification signal is used for determining whether a result obtained by verifying whether two items of ECG training data belong to the same entity through the nonlinear mapping of the layers is valid");
forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects (paragraph [0055], "A characteristic of a passing frequency band of the band pass filter is determined in a process of training a neural network model used for extracting a semantic feature of the ECG"; paragraph [0065], "the training device acquires ECG training data having various frequency bands using a plurality of band pass filters corresponding to different passbands"; paragraph [0095]); and
applying at least one machine learning process to the training dataset to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification (paragraph [0059], "the ECG authentication apparatus extracts the semantic feature of the ECG signal using the neural network model. The neural network model is a feature extracting model previously trained based on training data"; paragraph [0073], "The deep training method indicates a machine learning algorithm").
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg with the teachings of Liu so that the method comprises forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects; and applying at least one machine learning process to the training dataset, to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification, because doing so improves the performance of the model (Liu, paragraphs [0071]-[0072]).
Liu does not explicitly disclose that the at least one nonlinear mapping comprises at least one of: rectification; exponentiation; and gain control.
However, Frady teaches methods of encoding and computing (Abstract) comprising applying at least one nonlinear mapping comprising exponentiation (page 2, "The function domain is encoded by exponentiating a fixed random base vector"; page 15, exponential map).
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg and Liu with the teachings of Frady so that the at least one nonlinear mapping comprises exponentiation, because doing so provides transparency and insight into the machine learning model's features and weights, which helps explain the decisions of a neural network in specific application settings (Frady, page 1, Introduction, first paragraph; page 2, "VFA is fully transparent. Vectors can represent individual data points as well as elements of a function space that is well-defined as a reproducing kernel Hilbert space. The domain of the functions can encode continuous-valued quantities in data, such as position, time or wavelength"; page 21, "Changing the input encoding from random projection methods in conventional reservoir computing to KLPEs will yield a new class of recurrent networks for computing in a transparent fashion in a well-defined function space").
Regarding claim 6, the method of claim 1 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the subject classification comprises at least one characteristic selected from the group consisting of: age; gender; race; a physiological condition; a mental condition; a health condition; and any combination thereof (paragraph [0048]).
Regarding claim 7, the method of claim 1 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses simultaneously monitoring or identifying multiple subjects in a location using source separation techniques (paragraph [0092], "The processing means is configured to source separating by component analysis a mixture of detected signals and further to recover and extract the desired and preselected component signal(s) from a mixture of signals"; paragraph [0115], "the system may be located at any desired location and simultaneously spatially resolve, distinguish, and determine the physiological parameters arising from multiple subjects by way of utilizing differing range, angular azimuthal, and/or angular elevation information to each uniquely distinguished subject detected").
Regarding claim 8, Steinberg discloses a system for biometric identification (Fig. 1, paragraphs [0100], [0108]), the system comprising:
a cardiac signal detector (Fig. 1, paragraph [0081], receiver 120), configured to
receive a reflection radar signal reflected from a body tissue of a monitored subject, wherein the reflection radar signal is in the frequency range of 0.03 to 3 THz (paragraph [0080], "a remote portable non-contact detection system ... one or more reception means for receiving the sub-THz and THz signal of the subject. The received sub-THz and THz signals being a reflection of the sub-THz and THz signal from subject thereby"; paragraph [0088] ,"electromagnetic waves within the ITU-designated band of frequencies from 0.03 to 3 terahertz"), and
to derive a cardiac ballistocardiogram (BCG) signal from the reflection THz radar signal (paragraph [0090], "The microprocessor is further configured to perform analysis, calculation, data processing, automated reasoning, storing and/or processing the received sub-THz and THz signals and detect at least one physiological parameter"; paragraph [0088], "The term ‘Physiological Parameters’ herein refers to any physiological indicator...such as...ballistocardiogram(BCG), BCG amplitude variability");
a cardiac signal processor (Fig. 1, paragraph [0082], data processing means 150; paragraph [0090]), configured to:
segment the derived BCG signal into a plurality of discrete temporal segments, each temporal segment being of a selected time duration (paragraph [0133], "folding or mirroring the signals and decimating selected portions of the folded signals and removing folded segments");
apply at least one linear mapping to each of the temporal segments to produce a heartbeat frequency encoding (paragraph [0133], bandpass filtering); and
identify a subject based on the BCG signal according to stored data (paragraphs [0100], [0108], [0113]); and
wherein the system is further configured to, in response to classifying the heartbeat frequency encoding as belonging to the reference subject, perform at least one action selected from: associating collected cardiac signal data of the monitored subject with stored information of the reference subject (paragraph [0091], "The microprocessor may further detect, compare and provide interpretation of the received signal indicating of a change in subject's health condition, based on the received signal information and/or stored information"; paragraphs [0019], [0109]); or generating a notification based on the identification of the monitored subject (paragraph [0107], "the system may further export an output associated with the health status selected from the group consisting of an alert, an indication flag, an activation instruction of electronic device, electronic message or any combination thereof"; paragraphs [0015], [0021], [0042], [0051], [0124]).
Steinberg does not explicitly disclose segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat; applying at least one nonlinear mapping comprising at least one of: rectification; exponentiation; and gain control, to the temporal segments to produce a heartbeat frequency encoding; nor applying a trained machine learning subject classification model to the heartbeat frequency encoding to classify the heartbeat frequency encoding as belonging to a reference subject of a reference subject group if a matching classification is obtained or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no machine classification is obtained.
However, Liu teaches a method for ECG authentication (Abstract) comprising:
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat (Fig. 2, paragraph [0055], step 220, "The preprocessing includes detection of a fiducial point and acquirement of a data segment"; paragraph [0067], acquirement of data segments; paragraph [0100], "The data segments acquired based on the point R includes overall information associated with a single heartbeat");
applying at least one linear mapping comprising a plurality of bandpass filters to each of the temporal segments (Fig. 2, paragraph [0055], step 220, "the ECG authentication apparatus filters the ECG signal using a band pass filter configured to pass a predefined frequency band"; paragraphs [0017]-[0018], [0065], [0095]-[0097]);
applying at least one nonlinear mapping to the temporal segments to produce a heartbeat frequency encoding (paragraph [0073], "The deep training method indicates a machine learning algorithm for attempting a high-level abstraction by combining various non-linear transformation schemes"; paragraph [0080], "The identification signal is used for determining whether a result of identification performed on an entity of a type corresponding to the ECG training data through a nonlinear mapping of layers is valid. The verification signal is used for determining whether a result obtained by verifying whether two items of ECG training data belong to the same entity through the nonlinear mapping of the layers is valid");
applying at least one machine learning model for subject classification on the heartbeat frequency encoding during an identification stage to classify the heartbeat frequency encoding as belonging to a reference subject if a matching classification is obtained, or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no matching classification is obtained (Fig. 2, paragraph [0061], step 240, "The ECG authentication apparatus calculates a similarity between the semantic feature and a predefined registered feature or a reference feature corresponding to a target to be compared with the semantic feature. The authentication apparatus determines an authentication result to be a success in authentication or a fail in authentication based on a comparison result of the calculated similarity and a threshold").
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg with the teachings of Liu so that the method comprises segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat; applying at least one nonlinear mapping, to the temporal segments to produce a heartbeat frequency encoding; nor applying a trained machine learning subject classification model to the heartbeat frequency encoding to classify the heartbeat frequency encoding as belonging to a reference subject of a reference subject group if a matching classification is obtained or to determine that the heartbeat frequency encoding belongs to a non-reference subject if no machine classification is obtained, because doing so provides a method of authentication using markers that are not easily stolen, lost, or forged (Liu, paragraph [0005]).
Liu does not explicitly disclose that the at least one nonlinear mapping comprises at least one of: rectification; exponentiation; and gain control.
However, Frady teaches methods of encoding and computing (Abstract) comprising applying at least one nonlinear mapping comprising exponentiation (page 2, "The function domain is encoded by exponentiating a fixed random base vector"; page 15, exponential map).
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg and Liu with the teachings of Frady so that the at least one nonlinear mapping comprises exponentiation, because doing so provides transparency and insight into the machine learning model's features and weights, which helps explain the decisions of a neural network in specific application settings (Frady, page 1, Introduction, first paragraph; page 2, "VFA is fully transparent. Vectors can represent individual data points as well as elements of a function space that is well-defined as a reproducing kernel Hilbert space. The domain of the functions can encode continuous-valued quantities in data, such as position, time or wavelength"; page 21, "Changing the input encoding from random projection methods in conventional reservoir computing to KLPEs will yield a new class of recurrent networks for computing in a transparent fashion in a well-defined function space").
Regarding claim 9, the system of claim 8 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the machine learning subject classification model is generated during a training stage comprising:
for each of a plurality of reference subjects (paragraphs [0109], [0115]),
receiving a reflection THz radar signal reflected from a body tissue of a monitored subject using contactless detection (paragraph [0080], "a remote portable non-contact detection system ... one or more reception means for receiving the sub-THz and THz signal of the subject. The received sub-THz and THz signals being a reflection of the sub-THz and THz signal from subject thereby");
deriving a cardiac ballistocardiogram (BCG) signal from the reflection THz radar signal (paragraph [0090], "The microprocessor is further configured to perform analysis, calculation, data processing, automated reasoning, storing and/or processing the received sub-THz and THz signals and detect at least one physiological parameter"; paragraph [0088], "The term ‘Physiological Parameters’ herein refers to any physiological indicator...such as...ballistocardiogram(BCG), BCG amplitude variability"); and
the cardiac signal processor is configured
to segment the derived BCG signal into a plurality of discrete temporal segments, each temporal segment being of a selected time duration (paragraph [0133], "folding or mirroring the signals and decimating selected portions of the folded signals and removing folded segments");
to apply at least one linear mapping comprising a plurality of bandpass filters to each of the temporal segments (paragraph [0133], bandpass filtering; page 19, line 15 of the instant specification discloses that linear mapping can include band pass filtering); and
to assign the heartbeat frequency encoding to an identification label relating to the reference subject (paragraphs [0100], [0108], [0113]).
Steinberg does not explicitly disclose applying at least one nonlinear mapping comprising at least one of: rectification; exponentiation; and gain control, to the temporal segments to produce a heartbeat frequency encoding; nor forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects; and applying at least one machine learning process to the training dataset, to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification.
However, Liu further teaches:
segmenting the derived BCG signal into a plurality of discrete temporal segments, each temporal segment representing a duration of an individual heartbeat (Fig. 2, paragraph [0055], step 220, "The preprocessing includes detection of a fiducial point and acquirement of a data segment"; paragraph [0067], acquirement of data segments; paragraph [0100], "The data segments acquired based on the point R includes overall information associated with a single heartbeat");
applying at least one nonlinear mapping to the temporal segments to produce a heartbeat frequency encoding (paragraph [0073], "The deep training method indicates a machine learning algorithm for attempting a high-level abstraction by combining various non-linear transformation schemes"; paragraph [0080], "The identification signal is used for determining whether a result of identification performed on an entity of a type corresponding to the ECG training data through a nonlinear mapping of layers is valid. The verification signal is used for determining whether a result obtained by verifying whether two items of ECG training data belong to the same entity through the nonlinear mapping of the layers is valid");
forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects (paragraph [0055], "A characteristic of a passing frequency band of the band pass filter is determined in a process of training a neural network model used for extracting a semantic feature of the ECG"; paragraph [0065], "the training device acquires ECG training data having various frequency bands using a plurality of band pass filters corresponding to different passbands"; paragraph [0095]); and
applying at least one machine learning process to the training dataset to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification (paragraph [0059], "the ECG authentication apparatus extracts the semantic feature of the ECG signal using the neural network model. The neural network model is a feature extracting model previously trained based on training data"; paragraph [0073], "The deep training method indicates a machine learning algorithm").
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg with the teachings of Liu so that the method comprises forming a training dataset comprising a plurality of heartbeat frequency encodings obtained from the plurality of reference subjects; and applying at least one machine learning process to the training dataset, to identify classification profiles and patterns of the reference subjects for generating at least one predictive model for predicting a subject classification, because doing so improves the performance of the model (Liu, paragraphs [0071]-[0072]).
Liu does not explicitly disclose that the at least one nonlinear mapping comprises at least one of: rectification; exponentiation; and gain control.
However, Frady teaches methods of encoding and computing (Abstract) comprising applying at least one nonlinear mapping comprising exponentiation (page 2, "The function domain is encoded by exponentiating a fixed random base vector"; page 15, exponential map).
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Steinberg and Liu with the teachings of Frady so that the at least one nonlinear mapping comprises exponentiation, because doing so provides transparency and insight into the machine learning model's features and weights, which helps explain the decisions of a neural network in specific application settings (Frady, page 1, Introduction, first paragraph; page 2, "VFA is fully transparent. Vectors can represent individual data points as well as elements of a function space that is well-defined as a reproducing kernel Hilbert space. The domain of the functions can encode continuous-valued quantities in data, such as position, time or wavelength"; page 21, "Changing the input encoding from random projection methods in conventional reservoir computing to KLPEs will yield a new class of recurrent networks for computing in a transparent fashion in a well-defined function space").
Regarding claim 13, the system of claim 8 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the subject classification comprises at least one characteristic selected from the group consisting of: age; gender; race; a physiological condition; a mental condition; a health condition; and any combination thereof (paragraph [0048]).
Regarding claim 14, the system of claim 8 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses simultaneously monitoring or identifying multiple subjects in a location using source separation techniques (paragraph [0092], "The processing means is configured to source separating by component analysis a mixture of detected signals and further to recover and extract the desired and preselected component signal(s) from a mixture of signals"; paragraph [0115], "the system may be located at any desired location and simultaneously spatially resolve, distinguish, and determine the physiological parameters arising from multiple subjects by way of utilizing differing range, angular azimuthal, and/or angular elevation information to each uniquely distinguished subject detected").
Regarding claim 15, the method of claim 1 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the reflection radar signal has a frequency in the range of 77 GHz to 160 GHz (paragraph [0103], "the radar system may comprise silicon transceivers operating at a range...of about 77 GHz to 160 GHz").
Regarding claim 16, the system of claim 8 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the reflection radar signal has a frequency in the range of 77 GHz to 160 GHz (paragraph [0103], "the radar system may comprise silicon transceivers operating at a range...of about 77 GHz to 160 GHz").
Regarding claim 17, the method of claim 1 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the reflection radar signal is a frequency-modulated continuous wave (FMCW) radar signal (paragraph [0093], "schemas of the radar system as a single-frequency full duplex which may be continuous wave (CW) or frequency-modulated continuous wave (FMCW) radar system transceiver").
Regarding claim 18, the system of claim 8 is obvious over Steinberg, Liu, and Frady, as explained above. Steinberg further discloses that the reflection radar signal is a frequency-modulated continuous wave (FMCW) radar signal (paragraph [0093], "schemas of the radar system as a single-frequency full duplex which may be continuous wave (CW) or frequency-modulated continuous wave (FMCW) radar system transceiver").
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CHRISTINE SISON whose telephone number is (703)756-4661. The examiner can normally be reached 8 am - 5 pm PT, Mon - Fri.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jennifer McDonald can be reached at (571) 270-3061. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/CHRISTINE SISON/Examiner, Art Unit 3796
/PAMELA M. BAYS/Primary Examiner, Art Unit 3796
Prosecution Insights