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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 2/20/2025 has been entered.
Response to Arguments
Applicant's arguments filed 2/20/2025 regarding the objections to Claims 1 and 11 have been fully considered and are persuasive. Accordingly, the objections have been withdrawn.
Applicant’s arguments regarding the rejection of Claims 1, 9, 10, 11, 12, 13 and 16, and Claims 3, 6-8 and 14-15 by dependency under 35 USC 112(a) have been fully considered and are persuasive. Claims 1, 9, 10, 11, 12, 13 and 16 have been amended sufficiently to overcome the rejection. Accordingly, the rejections have been withdrawn.
It is noted that Claim 16 as amended is rejected again under 35 USC 112(a), but for different reasons. Applicant states that support for Claim 16 as amended can be found at Para. [0039], which states “In this manner, the leaning unit 110 of the embodiment may perform learning using the teacher data in accordance with the usage application of the measurement device 20 in which the learned model is provided.” It is not clear why Applicant believes that Para. [0039] supports a limitation which pertains to the device being moveable by the user.
Applicant’s arguments regarding the rejection of Claim 16 under 35 USC 112(b) have been fully considered and are persuasive. The Examiner agrees that the amendments to Claim 16 are sufficient to overcome the rejection thereto. Accordingly, the rejection is withdrawn.
Applicant’s arguments regarding the rejection of Claims 1, 3 and 6-16 under 35 USC 101 have been fully considered but are not persuasive. Claims 1, 3 and 6-16 recite abstract ideas because training a machine learning algorithm is a mathematical process. Contrary to Applicant’s arguments, the recited measurement devices are additional elements, and do not impact whether the machine learning algorithm is a mathematical calculation.
However, Claims 1, 3 and 6-16 as amended now recite patent eligible subject matter because the recited first and second “measurement systems” sufficiently integrate the mathematical calculation of training a machine learning algorithm into a practical application. The invention of Claims 1, 3 and 6-16 represents an improvement to the technology of sensing cardiac information of a user operating a vehicle. See MPEP 2106.04(d)(1) and 2106.05(a); see also Example 47, Claim 3 of the July 2024 Subject Eligibility Guidance. The Present Specification makes clear that this improvement is provided by the reduction of noise of in sensed data. See Present Specification at Para. [0005] (“…in view of the above-described problem, the present invention aims at providing a mechanism that is capable of acquiring vital data less affected by noises more efficiently.”). Consistent with MPEP 2106.05(a), (1) the claims cover a particular solution to a problem or a particular way to achieve a desired outcome and (2) the improvement is provided by one or more additional elements. Claim 1 recites “wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform.” Claims 9-13 recite a similar limitation. The difference in the amount of noise is based on the positioning of the recited first and second measurement systems. The machine learning algorithm removes noise from the first waveform based on the less noisy second waveform. This reduction of noise in the improvement. The positioning of the recited first and second measurement systems allows the machine learning algorithm to implement the improvement. Consequently, the recited first and second measurement systems integrate the recited abstract idea into a practical application. Accordingly, the rejection of Claims 1, 3 and 6-16 under 35 USC 101 is withdrawn.
Applicant’s arguments regarding the rejection of Independent Claims 1 and 9-10 under 35 USC 103 as unpatentable over US 2019/0183431 A1 to Attia et al. (“Attia”) in view of US 2012/0302860 A1 to Volpe et al. (“Volpe”) and Independent Claims 11-13 under 35 USC 103 as unpatentable over Attia and Volpe in view of Non-Patent Literature: Tadeáš Bednár, Branko Babušiak, "Measurement of capacitive coupled ECG from the car seat," Transportation Research Procedia, Volume 40, 2019, Pages 1260-1265 (“Bednár”) have been fully considered and are persuasive to the extent the Examiner agrees that none of Attia, Volpe and Bednar either alone or in combination teach the positioning of the sensors of the recited first and second measurement systems that in the manner of Claims 1 and 9-13 as amended. Accordingly, the rejection is withdrawn. However, upon further search and consideration, new grounds of rejection is made in view of US 2019/0082993 A1.
Applicant’s arguments on Pg. 22-23 of Applicant’s Remarks regarding the teachings of Attia have been fully considered but are not persuasive. Applicant argues that Attia does not teach the Claim 1, 9 and 10 limitation “performs learning on an output of presence probability data indicating a presence probability of a R wave in the first electrocardiographic waveform…” because “an ordinary artisan would reasonably understand that a presence probability of a R wave being present (i.e., a percentage value) resulting in a binary determination is patentably distinguished from a measure amplitude of an R wave that is known to be present (i.e., a value of magnitude)” This position is inconsistent with the Present Specification, which states at Para. [0044] “Although FIG. 4 and FIG. 5 exemplify the case in which the presence probability data is of two values of 0 (absent) or 1 (present), the presence probability data of the embodiment may be of three or more values.” (emphasis added). The Present Specification allows for both a binary “present or not present” and a particular percentage. Attia’s R-wave is present: therefore its presence probability data is 1 (present). Applicant’s arguments are not persuasive.
Applicant’s arguments regarding dependent Claims 3, 6-8 and 14-16 are based on Applicant’s arguments with respect to the Independent Claims from which each depends. Applicant’s arguments have been fully considered and are persuasive to the extent explained above. Accordingly, the rejection is withdrawn. However, upon further search and consideration, new grounds of rejection is made in view of US 2019/0082993 A1.
Claim Objections
Claim 1 is objected to because of the following informalities: Claim 1 recites “…wherein the second cardiographic includes a lesser amount…” at Ln. 10, but should recite --…wherein the second cardiographic waveform includes a lesser amount…--. Appropriate correction is required.
Claims 9-13 each contain the same informality, and are objected to for the same reasons.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1, 9-13 and 16, and Claims 3, 6-8 and 14-15 by dependency, are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Regarding Independent Claim 1:
Claim 1 recites “generate a modified learning algorithm based on the learning performed on the first electrocardiographic waveform in view of the second electrocardiographic waveform, generate a third electrocardiographic waveform using the modified learning algorithm, wherein the modified learning algorithm removes the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform.” The Present Specification does not provide support for generating such a “modified learning algorithm,” for generating a third electrocardiographic waveform “using the modified learning algorithm,” or for “the modified learning algorithm remov[ing] the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform.” The term “algorithm” appears only once in the Present Specification at Para. [0014], where it is stated that “The learning unit 110 performs learning using an algorithm such as a neural network or a support vector machine (SVM), for example.” No mention is made of modifying the algorithm, nor is any mention made of a modified algorithm. Instead, the Present Specification suggests that the recited machine learning algorithm itself generates the recited third electrocardiographic waveform using as its input the first and second electrocardiographic waveforms (see Paras. [0040] through [0042] of the Present Specification).
Claim 1 recites “wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform.” The Present Specification does not provide support for this limitation. The Present Specification states at Paras. [0006], [0007], [0008], [0009], [0013], [0020], [0024], [0027] and [0041] that the second measurement system is less affected by noise, but does not state anywhere that the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform. These two statements are different, as evidenced by the separate Claim 1 limitation “the second measurement system being less affected by noise than the first measurement system.”
Claims 9-13 contain similar limitations, which is not supported by the Specification for the same reasons.
Regarding Claim 16, Claim 16 recites “wherein an the device is configured to allow a movement of the device by the user.” Applicant states that support for Claim 16 as amended can be found at Para. [0039], which states “In this manner, the leaning unit 110 of the embodiment may perform learning using the teacher data in accordance with the usage application of the measurement device 20 in which the learned model is provided.” It is not clear why Applicant believes that Para. [0039] supports a limitation which pertains to the device being moveable by the user. The Present Specification does not provide support for the device being configured to allow a movement of the device by the user.
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1 and 9-13, and Claims 3, 6-8 and 14-16 by dependency, are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding Independent Claim 1:
Claim 1 recites “wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform.” It is unclear in what sense “the noise” (i.e., the “noise caused by at least a movement of the vehicle or the user…” as defined previously) can have two different amounts. Any particular movement that causes noises causes only one amount of noise, although two different measurement systems may be differently influenced by noise (for example as described at Paras. [0006], [0007], [0008], [0009], [0013], [0020], [0024], [0027] and [0041] of the Present Specification).
For purposes of this Office Action, the term “wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform” is being interpreted to mean that the second electrocardiographic waveform reflects the lesser influence of noise on the second measurement system.
Claim 1 recites “…generate a modified learning algorithm….” It is unclear in what sense the learning algorithm is “modified.” For example, modification of a learning algorithm might entail generating an entirely new algorithm, altering the weighting of certain parameters used by the algorithm, or something else. The Present Specification provides no guidance with respect to the term. The scope of the claim thus cannot be discerned.
For purposes of this Office Action, the term “generate a modified learning algorithm” is being interpreted to mean that any aspect of the algorithm is different from a previous iteration, be that difference in weighting, input, output, or something else.
Claim 1 recites “wherein the first electrocardiographic waveform being output based on first sensor data of the user acquired from the first measurement system as learning data.” It is grammatically unclear what sense this limitation limits the claim. For example, it is unclear whether the “first electrocardiographic waveform” is output based on learning data, whether the “first electrocardiographic waveform” is used as learning data, or something else.
For purposes of this Office Action, the above limitation is being interpreted to mean that “the first electrocardiographic waveform” is used as learning data, which waveform (as recited previously in the claim) is output based on first sensor data that is acquired from the first measurement system.
Claim 1 recites “the second electrocardiographic waveform being output based on second sensor data of the user acquired from the second measurement system as teacher data.” It is grammatically unclear what sense this limitation limits the claim. For example, it is unclear whether the “second electrocardiographic waveform” is output based on teacher data, or whether the “second electrocardiographic waveform” is used as teacher data, or something else.
For purposes of this Office Action, the above limitation is being interpreted to mean that “the second electrocardiographic waveform” is used as teacher data, which waveform (as recited previously in the claim) is output based on second sensor data that is acquired from the second measurement system.
Claim 1 recites “wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform.” However, the subject first and second electrocardiographic waveforms are obtained during use of the device, and the amount of noise they contain is a result of that use. The Present Specification states at Paras. [0006], [0007], [0008], [0009], [0013], [0020], [0024], [0027] and [0041] that the second measurement system is less affected by noise, but the mere fact that a system is less affected by noise does not necessarily mean that less noise will be included in its measurements: the amount of noise in the measurement is additionally influenced by the amount of noise imparted by use. It is unclear in what sense a result of using the device (i.e., the amount of noise measured during use) impacts the structure of the claimed device.
Regarding Claims 9-13, Claims 9-13 recite similar limitations to those of Claim 1, which are indefinite for the same reasons.
Claim Rejections - 35 USC § 103
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.
Claims 1, 3, 6 and 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over US 2019/0082993 A1 to Choi et al. (“Choi”) in view of previously cited US 2019/0183431 A1 to Attia et al. (“Attia”) and previously cited US 2012/0302860 A1 to Volpe et al. (“Volpe”) as evidenced by Roy, A; "An Introduction to Gradient Descent and Backpropagation;" published on Medium.com, https://medium.com/data-science/an-introduction-to-gradient-descent-and-backpropagation-81648bdb19b2 on June 14, 2020; accessed 10/28/2025 (“Roy”).
Regarding Independent Claim 1, Choi teaches:
A learning system, comprising: (Title, “Electrocardiogram measurement device for vehicle, system including the same, and method therefor;” Para. [0011]);
The preamble recitation “learning” is being interpreted as an intended use of the recited system. See MPEP 2111.02.
Choi’s Para. [0011] describes evaluating the quality of a captured ECG signal and making adjustments accordingly. This evaluation is such learning as recited.
a first measurement system including a first plurality of sensors disposed on a vehicle, (Para. [0057], “In an exemplary embodiment having electrodes installed in a steering wheel, a change gear, a radio volume adjuster, a window, a driver's seat door armrest, an ECG of the driver may be more accurately measured in various driving situations.”);
Choi’s electrodes installed in “a driver’s seat door armrest” are such a “first measurement system” as claimed.
the first plurality of sensors configured to measure and output a first electrocardiographic waveform of a user of the vehicle while the vehicle is being driven by the user, (Para. [0056], “An exemplary embodiment of the present disclosure may be configured to measure an ECG although the first or second hand of the driver are disposed at various positions, by installing a plurality of electrodes in various positions in a vehicle.”);
wherein the first electrocardiographic waveform includes noise caused by at least a movement of the vehicle or the user; (Para. [0014], “In addition, the signal quality evaluator may be configured to detect strength and a noise component of the ECG signal and evaluate the quality of the ECG signal based on the detected strength and noise component of the ECG signal. In an exemplary embodiment, the electrode may be disposed in of a steering wheel, a gear change lever, a window button, a door armrest, a radio volume adjuster, an armrest, a handle, a horn, a turn signal level, a glove compartment, or a seat belt.”);
a second measurement system including a second plurality of sensors configured to be disposed on a skin of the user of the vehicle, (Para. [0057], “In an exemplary embodiment having electrodes installed in a steering wheel, a change gear, a radio volume adjuster, a window, a driver's seat door armrest, an ECG of the driver may be more accurately measured in various driving situations.”);
Choi’s electrodes “installed in a steering wheel” are “configured to be disposed on a skin of the user of the vehicle” in the manner claimed, as they perform ECG measurement when the user’s hand is in contact with the steering wheel (see Choi at Para. [0040]).
the second plurality of sensors being configured to measure and output a second electrocardiographic waveform of the user while the vehicle is being driven by the user, (Para. [0016], “… an ECG measurement method for a vehicle may include when recognizing contact of a body of a driver with a first electrode among an electrode in contact with the body of the driver, outputting a first electrode signal from the first electrode, when recognizing contact of a body of the driver with a second electrode, outputting a second electrode signal from the second electrode may be output, an ECG signal by differentially amplifying the first electrode signal and the second electrode signal,…”);
a processor configured to execute a … algorithm (Para. [0070], “Thus, the operations of the methods or algorithms described in connection with the exemplary embodiments disclosed in the specification may be directly implemented with a hardware module, a software module, or two combinations thereof, executed by the processor 1100.”);
Choi does not disclose:
wherein the second electrocardiographic includes a lesser amount of the noise than the first electrocardiographic waveform;
a processor configured to execute a machine learning algorithm
that: performs learning on the first electrocardiographic waveform of the user output by the first measurement system in view the second electrocardiographic waveform output by the second measurement system,
generate a modified learning algorithm based on the learning performed on the first electrocardiographic waveform in view of the second electrocardiographic waveform,
generate a third electrocardiographic waveform using the modified learning algorithm,
wherein the modified learning algorithm removes the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform,
and acquire one or more physical indices of the user based on the third electrocardiographic waveform generated using the modified learning algorithm,
wherein the first electrocardiographic waveform being output based on first sensor data of the user acquired from the first measurement system as learning data,
the second electrocardiographic waveform being output based on second sensor data of the user acquired from the second measurement system as teacher data,
the learning is performed using the teacher data based on the second sensor data of the user acquired from the subject by a second measurement system in a same period as an acquisition period of the first sensor data,
the second measurement system being less affected by noises than the first measurement system,
each of the first, second and third electrocardiographic waveforms vital data includes data related to a cardiac activity,
the machine learning algorithm learns output of data related to the cardiac activity of the user,
with the use of the first electrocardiographic waveform acquired by the first measurement system as the learning data and the teacher data based on the second electrocardiographic waveform acquired by the second measurement system in a same period as an acquisition period of the first electrocardiographic waveform,
and the machine learning algorithm performs learning on an output of presence probability data indicating a presence probability of a R wave in the first electrocardiographic waveform using presence probability data that indicates the presence probability of the R wave in the second electrocardiographic waveform as the teacher data, and calculates a heartbeat cycle defined by an interval between R waves based on the present probability of the R wave in the first electrocardiographic waveform
Attia describes “Predicting transient ischemic events using ecg data” (Title). Attia is analogous art.
Attia teaches:
a processor configured to execute a machine learning algorithm (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event.);
that: performs learning on the first electrocardiographic waveform of the user output by the first measurement system in view the second electrocardiographic waveform output by the second measurement system, (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event. For example, a system employing a machine-learning model can be trained with a large training set of ECG data for known patients with and without ischemic events (e.g., acute stroke events) and can be used to process a recording of ECG data from a patient to generate a prediction indicating a likelihood that the patient will experience a stroke;” Para. [0061], “An ECG recording is obtained from the patient (354). … The system generates neural network inputs based on the ECG recording and the auxiliary data, if available (358). The ischemic event detection neural network processes the neural network inputs to generate the ischemic event detection (360);” Para. [0057], “In alternative embodiments where the ischemic event prediction/detection neural network 108 operates as an ischemic event detection neural network 108, the training neural network subsystem 206 for the ischemic event detection neural network 108 is trained on ECG training data that includes a plurality of ECG training samples;” Para.[0058]);
generate a modified learning algorithm based on the learning performed on the first electrocardiographic waveform in view of the second electrocardiographic waveform, (Para. [0061], “The ischemic event detection neural network processes the neural network inputs to generate the ischemic event detection (360).”);
generate a third electrocardiographic waveform using the modified learning algorithm, (Para. [0059], “After training is complete, the training system 200 can provide a final set of parameter values 218 to the system 100 for use in making ischemic event predictions or detections 120. The training system 200 can provide the final set of model parameter values 218 by a wired or wireless connection to the system 100 and neural network 108, for example.”);
Attia’s “final set of parameter values” is such a “third electrocardiographic waveform” as claimed, and is generated using Attia’s modified learning algorithm.
wherein the modified learning algorithm removes the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform, (Para. [0004], “Machine-learning techniques such as backpropagation of errors with gradient descent can be used to train the model.”)
Backpropagation of errors with gradient descent removes noise. See Roy, A; "An Introduction to Gradient Descent and Backpropagation;" published on Medium.com, https://medium.com/data-science/an-introduction-to-gradient-descent-and-backpropagation-81648bdb19b2 on June 14, 2020; accessed 10/28/2025.
and acquire one or more physical indices of the user based on the third electrocardiographic waveform generated using the modified learning algorithm, (Para. [0046], “A number of morphological features that may be employed for ischemic event prediction/diagnosis are labeled in FIG. 6, such as a duration of the QRS-complex, an amplitude of the P-wave, R-wave, or T-wave, an area of the P-wave, QRS-complex, or T-wave, slopes of any of the waves, distances between the waves, and centers-of-gravity of the waves.”);
See Paras. [0033] through [0036] of the Present Specification, describing R-waves as such a “physical index” as claimed.
wherein the first electrocardiographic waveform being output based on first sensor data of the user acquired from the first measurement system as learning data, (Para. [0003], Para. [0061]);
As explained above, this limitation is being interpreted to mean that “the first electrocardiographic waveform” is used as learning data, which waveform (as recited previously in the claim) is output based on first sensor data that is acquired from the first measurement system.
the second electrocardiographic waveform being output based on second sensor data of the user acquired from the second measurement system as teacher data, (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event. For example, a system employing a machine-learning model can be trained with a large training set of ECG data for known patients with and without ischemic events (e.g., acute stroke events) and can be used to process a recording of ECG data from a patient to generate a prediction indicating a likelihood that the patient will experience a stroke;” Para. [0057], Para. [0058])
As explained above, this limitation is being interpreted to mean that “the second electrocardiographic waveform” is used as teacher data, which waveform (as recited previously in the claim) is output based on second sensor data that is acquired from the first measurement system.
each of the first, second and third electrocardiographic waveforms vital data includes data related to a cardiac activity, (Para. [0003]; Para. [0057], Para. [0058])
Attia’s first, second and third electrocardiographic waveforms are ECG data, and thus include data related to a cardiac activity.
the machine learning algorithm learns output of data related to the cardiac activity of the user, (Para. [0003]; Para. [0057], Para. [0058]);
Attia’s machine learning algorithm learns ECG data, and thus learns data related to a cardiac activity.
and the machine learning algorithm performs learning on an output of presence probability data indicating a presence probability of a R wave in the first electrocardiographic waveform using presence probability data that indicates the presence probability of the R wave in the second electrocardiographic waveform as the teacher data, and calculates a heartbeat cycle defined by an interval between R waves based on the present probability of the R wave in the first electrocardiographic waveform (Para. [0050], “The prediction 120 can be expressed as a probability or confidence score representing a probability or confidence that the patient 102 will experience an ischemic event within a pre-defined time from a time when the ECG recording was obtained. In some implementations, the prediction 120 is expressed as a selection of a particular classification from multiple possible classifications that represents a most likely condition of the patient 102;” Para. [0046], “A number of morphological features that may be employed for ischemic event prediction/diagnosis are labeled in FIG. 6, such as a duration of the QRS-complex, an amplitude of the P-wave, R-wave, or T-wave, an area of the P-wave, QRS-complex, or T-wave, slopes of any of the waves, distances between the waves, and centers-of-gravity of the waves.”).
It would have been obvious for a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the device of Choi with the teachings of Attia (i.e., to modify the device of Choi so that its processor employs such a machine learning algorithm as that taught by Attia) in order to facilitate detection of ischemic events (Attia at Abstract).
Volpe describes “detection of cardiac function and the treatment of cardiac conditions in an ambulatory medical device” (Para. [0003]). Volpe is thus analogous art.
Volpe teaches:
wherein the second electrocardiographic waveform includes a lesser amount of the noise than the first electrocardiographic waveform; (Para. [0032], “the wearable medical device resolves conflicts between these two channels in favor of a previously identified, preferred channel. Further, in some examples, the set of preference information is automatically configured by the wearable medical device and continuously adjusted during its operation. For example, the set of preference information may be adjusted based on the current health of the patient, the activity of the patient, and the present locations of the electrodes relative to the patient's body.”);
the learning is performed using the teacher data based on the second sensor data of the user acquired from the subject by a second measurement system in a same period as an acquisition period of the first sensor data, (Para. [0010]; Fig. 2, “Receive an ECG Signal from Each Channel”);
the second measurement system being less affected by noises than the first measurement system, (Para. [0032]; Para. [0006]);
with the use of the first electrocardiographic waveform acquired by the first measurement system as the learning data and the teacher data based on the second electrocardiographic waveform acquired by the second measurement system in a same period as an acquisition period of the first electrocardiographic waveform (Para. [0010]; Fig. 2, “Receive an ECG Signal from Each Channel”).
It would have been obvious for a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the device of combined Choi and Attia with the teachings of Volpe (i.e., to use Choi’s data that is less effected by noise in the manner of Volpe) in order to increase the accuracy of acquired signals for subsequent interpretation by accounting for signal quality variations caused by movement (Volpe at Paras. [0005] through [0006]).
It would have been obvious for a person of ordinary skill in the art before the effective filing date of the claimed invention to modify the device of combined Choi and Attia with the teachings of Volpe (i.e., to obtain data from both of Choi’s sensors during the same time period in the manner of Volpe, and to subsequently use such simultaneously obtained data in the algorithm of Attia) in order to facilitate resolution of conflicts between interpretations of first ECG signal and the second ECG signal (Volpe at Para. [0010]).
Regarding Claim 3, the combination of Choi, Attia and Volpe renders obvious the entirety of Claim 1 as explained above.
Attia additionally discloses:
wherein the machine learning algorithm performs learning related to output of a third electrocardiographic waveform generated by removing noises from the first electrocardiographic waveform with the use of the second electrocardiographic waveform as teacher data. (Para. [0053], “The training system 200 includes a training neural network subsystem 206 that can implement the operations of each layer of a neural network that is designed to make an ischemic event prediction 120 or an ischemic event detection 120 from ECG recordings and, optionally, auxiliary information such as morphological features and patient profile data;” Para. [0059], “After training is complete, the training system 200 can provide a final set of parameter values 218 to the system 100 for use in making ischemic event predictions or detections 120. The training system 200 can provide the final set of model parameter values 218 by a wired or wireless connection to the system 100 and neural network 108, for example.”).
Attia’s disclosed “provid[ing] a final set of parameter values” is such “perform[ing] learning related to output of a third electrocardiographic waveform generated by removing noises from the first electrocardiographic waveform with the use of the second electrocardiographic waveform as teacher data” as claimed when the term is afforded its broadest reasonable interpretation.
Regarding Claim 6, the combination of Choi, Attia and Volpe renders obvious the entirety of Claim 1 as explained above.
Choi additionally teaches:
wherein the first measurement system is a system for acquiring an electrocardiographic waveform using at least two electrodes in contact with the user, and the second measurement system is a system for acquiring an electrocardiographic waveform using at least two electrodes attached on a skin of the user (Para. [0041], “Referring to FIG. 6, when a driver pushes a window button 620 with his or her first hand (e.g., left hand or right hand) to open a window 610 and holds a steering wheel 510 with his or her second hand (e.g., right hand or left hand), electrode signals may be output from an electrode 621 mounted on the window button 620 and an electrode 521 mounted on the steering wheel 510 and his or her ECG may be measured from a differential amplification value of the two electrode signals.”).
Regarding Claim 8, the combination of Choi, Attia and Volpe renders obvious the entirety of Claim 1 as explained above.
Choi additionally teaches:
wherein the user is a driver driving the vehicle (Para. [0009], “ The ECG measurement device for the vehicle may further include a contact sensor configured to detect that the body of the driver is in contact with the electrode…”).
Regarding Independent Claim 9, Choi teaches:
A learning method, comprising: (Title, “Electrocardiogram measurement device for vehicle, system including the same, and method therefor;” Para. [0011]);
Choi’s Para. [0011] describes evaluating the quality of a captured ECG signal and making adjustments accordingly. This evaluation is such learning as recited.
measuring and outputting, by a first measurement system including a first plurality of sensors disposed on a vehicle, a first electrocardiographic waveform of a user of the vehicle while the vehicle is being driven by the user, wherein the first electrocardiographic waveform includes noise caused by at least a movement of the vehicle or the user; (Para. [0057], “In an exemplary embodiment having electrodes installed in a steering wheel, a change gear, a radio volume adjuster, a window, a driver's seat door armrest, an ECG of the driver may be more accurately measured in various driving situations;” Para. [0056], “An exemplary embodiment of the present disclosure may be configured to measure an ECG although the first or second hand of the driver are disposed at various positions, by installing a plurality of electrodes in various positions in a vehicle;” Para. [0014], “In addition, the signal quality evaluator may be configured to detect strength and a noise component of the ECG signal and evaluate the quality of the ECG signal based on the detected strength and noise component of the ECG signal. In an exemplary embodiment, the electrode may be disposed in of a steering wheel, a gear change lever, a window button, a door armrest, a radio volume adjuster, an armrest, a handle, a horn, a turn signal level, a glove compartment, or a seat belt.”);
measuring and outputting, by a second measurement system including a second plurality of sensors configured to be disposed on a skin of the user of the vehicle, a second electrocardiographic waveform of the user while the vehicle is being driven by the user, (Para. [0057], “In an exemplary embodiment having electrodes installed in a steering wheel, a change gear, a radio volume adjuster, a window, a driver's seat door armrest, an ECG of the driver may be more accurately measured in various driving situations;” Para. [0016], “… an ECG measurement method for a vehicle may include when recognizing contact of a body of a driver with a first electrode among an electrode in contact with the body of the driver, outputting a first electrode signal from the first electrode, when recognizing contact of a body of the driver with a second electrode, outputting a second electrode signal from the second electrode may be output, an ECG signal by differentially amplifying the first electrode signal and the second electrode signal,…”);
performing, by a … algorithm executed by a processor (Para. [0070], “Thus, the operations of the methods or algorithms described in connection with the exemplary embodiments disclosed in the specification may be directly implemented with a hardware module, a software module, or two combinations thereof, executed by the processor 1100.”);
Choi does not disclose:
wherein the second electrocardiographic includes a lesser amount of the noise than the first electrocardiographic waveform;
performing, by a machine learning algorithm executed by a processor
learning on the first electrocardiographic waveform of the user vital data indicating life signs of a user output by the first measurement system in view the second electrocardiographic waveform output by the second measurement system;
generating a modified learning algorithm based on the learning performed on the first electrocardiographic waveform in view of the second electrocardiographic waveform;
generating a third electrocardiographic waveform using the modified learning algorithm, wherein the modified learning algorithm removes the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform;
and acquiring one or more physical indices of the user based on the third electrocardiographic waveform generated using the modified learning algorithm,
wherein the first electrocardiographic waveform being output based on first sensor data of the user acquired from the first measurement system as learning data,
the second electrocardiographic waveform being output based on second sensor data of the user acquired from the second measurement system as teacher data,
the learning is performed using the teacher data based on the second sensor data of the user acquired from the second measurement system in a same period as an acquisition period of the first sensor data,
the second measurement system being less affected by noises than the first measurement system,
each of the first, second and third electrocardiographic waveforms includes data related to a cardiac activity, the machine learning algorithm learns output of data related to the cardiac activity of the subject user,
with the use of the first electrocardiographic waveform acquired by the first measurement system as the learning data and the teacher data based on the second electrocardiographic waveform acquired by the second measurement system in a same period as an acquisition period of the first electrocardiographic waveform,
and the machine learning algorithm performs learning on an output of presence probability data indicating a presence probability of a R wave in the first electrocardiographic waveform using presence probability data that indicates the presence probability of the R wave in the second electrocardiographic waveform as the teacher data, and calculates a heartbeat cycle defined by an interval between R waves based on the present probability of the R wave in the first electrocardiographic waveform
Attia describes “Predicting transient ischemic events using ecg data” (Title). Attia is analogous art.
Attia teaches:
performing, by a machine learning algorithm executed by a processor (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event.);
learning on the first electrocardiographic waveform of the user vital data indicating life signs of a user output by the first measurement system in view the second electrocardiographic waveform output by the second measurement system; (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event. For example, a system employing a machine-learning model can be trained with a large training set of ECG data for known patients with and without ischemic events (e.g., acute stroke events) and can be used to process a recording of ECG data from a patient to generate a prediction indicating a likelihood that the patient will experience a stroke;” Para. [0061], “An ECG recording is obtained from the patient (354). … The system generates neural network inputs based on the ECG recording and the auxiliary data, if available (358). The ischemic event detection neural network processes the neural network inputs to generate the ischemic event detection (360);” Para. [0057], “In alternative embodiments where the ischemic event prediction/detection neural network 108 operates as an ischemic event detection neural network 108, the training neural network subsystem 206 for the ischemic event detection neural network 108 is trained on ECG training data that includes a plurality of ECG training samples;” Para.[0058]);
generating a modified learning algorithm based on the learning performed on the first electrocardiographic waveform in view of the second electrocardiographic waveform; (Para. [0061], “The ischemic event detection neural network processes the neural network inputs to generate the ischemic event detection (360).”);
generating a third electrocardiographic waveform using the modified learning algorithm, (Para. [0059], “After training is complete, the training system 200 can provide a final set of parameter values 218 to the system 100 for use in making ischemic event predictions or detections 120. The training system 200 can provide the final set of model parameter values 218 by a wired or wireless connection to the system 100 and neural network 108, for example.”);
Attia’s “final set of parameter values” is such a “third electrocardiographic waveform” as claimed, and is generated using Attia’s modified learning algorithm.
wherein the modified learning algorithm removes the noise included in the first electrocardiographic waveform to generate the third electrocardiographic waveform; (Para. [0004], “Machine-learning techniques such as backpropagation of errors with gradient descent can be used to train the model.”)
Backpropagation of errors with gradient descent removes noise. See Roy, A; "An Introduction to Gradient Descent and Backpropagation;" published on Medium.com, https://medium.com/data-science/an-introduction-to-gradient-descent-and-backpropagation-81648bdb19b2 on June 14, 2020; accessed 10/28/2025.
and acquiring one or more physical indices of the user based on the third electrocardiographic waveform generated using the modified learning algorithm, (Para. [0046], “A number of morphological features that may be employed for ischemic event prediction/diagnosis are labeled in FIG. 6, such as a duration of the QRS-complex, an amplitude of the P-wave, R-wave, or T-wave, an area of the P-wave, QRS-complex, or T-wave, slopes of any of the waves, distances between the waves, and centers-of-gravity of the waves.”);
See Paras. [0033] through [0036] of the Present Specification, describing R-waves as such a “physical index” as claimed.
wherein the first electrocardiographic waveform being output based on first sensor data of the user acquired from the first measurement system as learning data, (Para. [0003], Para. [0061]);
As explained above, this limitation is being interpreted to mean that “the first electrocardiographic waveform” is used as learning data, which waveform (as recited previously in the claim) is output based on first sensor data that is acquired from the first measurement system.
the second electrocardiographic waveform being output based on second sensor data of the user acquired from the second measurement system as teacher data, (Para. [0003], “A machine-learning model such as a neural network can be generated that allows for recognition of an ECG consistent with an ischemic event. For example, a system employing a machine-learning model can be trained with a large training set of ECG data for known patients with and without ischemic events (e.g., acute stroke events) and can be used to process a recording of ECG data from a patient to generate a prediction indicating a likelihood that the patient will experience a stroke;” Para. [0057], Para. [0058])
As explained above, this limitation is being interpreted to mean that “the second electrocardiographic waveform” is used as teacher data, which waveform (as recited previously in the claim) is output based on second sensor data that is acquired from the first measurement system.
each of the first, second and third electrocardiographic waveforms includes data related to a cardiac activity, (Para. [0003]; Para. [0057], Para. [0058])
Attia’s first, second and third electrocardiographic waveforms are ECG data, and thus include data related to a cardiac activity.
the machine learning algorithm learns output of data related to the cardiac activity of the subject user, (Para. [0003]; Para. [0057], Para. [0058]);
Attia’s machine learning algorithm learns ECG data, and thus learns data related to a cardiac activity.
and the machine learning algorithm performs learning on an output of presence probability data indicating a presence probability of a R wave in the first electrocardiographic waveform using presence probability data that indicates the presence probability of the R wave in the second electrocardiographic waveform as the teacher data, and calculates a heartbeat cycle defined by an interval between R waves based on the present probability of the R wave in the first electrocardiographic waveform(Para. [0050], “The prediction 120 can be expressed as a probability or confidence score representing a probability or confidence that the patient 102 will experience an ischemic event within a pre-defined time from a