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 Amendment
The amendment filed on 07/20/2026 has been entered. Claims 1 – 15, 17 – 23, 41 – 59, and 82 are pending. Claims 16, 24 – 40, and 60 – 81 are cancelled.
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.
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.
Claims 1 – 5, 7, 10, 11, 15, 17, 23, 41, 59, & 82 are rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30), as evidenced by Clifford (Clifford, G., et al., User Guide and Documentation for the MIMIC II Data Base, 24 Feb 2012).
In regard to claim 1, Vule discloses a system for estimating blood pressure comprising:
at least one sensor configured to measure at least one waveform related to blood pressure from a patient; Vule discloses a blood pressure assessment system with a plurality of sensors (FIG. 8, component 870) used to measure a biological signal, such as a PPG waveforms (FIG. 11, paragraph [0118]) and ECG waveforms (paragraph [0030]), which is used to isolate features related to blood pressure and estimate a physiological property, such as blood pressure, using a machine learning model (paragraphs [0007] - [0008]).
at least one processor (FIG. 8, component 805);
at least one computer-readable storage medium having encoded thereon executable instructions (paragraph [0009]) that, when executed by the at least one processor, cause the at least one processor to carry out a method comprising:
a) Measuring, using the at least one sensor, at least one waveform related to blood pressure from a patient. Vule discloses that sensors (FIG. 8, component 870) are used to measure a biological signal, such as a PPG waveforms (FIG. 11, paragraph [0118]) and ECG waveforms (paragraph [0030]).
b) receiving, from the at least one sensor, the at least one waveform and determining, from the at least one waveform, one or more features comprising at least one or more temporal features and or one or more morphology features; Vule discloses that a wrist worn PPG sensor (paragraph [0061]; FIG. 1, component 165) is used to collect PPG waveform data (paragraph [0066]) and extract a number of features from the collected blood flow pulse signal, such as shape of the pulse, a maximum amplitude, a minimum amplitude, a maximum derivative, a time difference between main and secondary peaks, and integral through the entire extraction time (paragraph [0067]; FIGs. 6A - 6C).
c) analyzing the one or more features using one or more trained models to determine one or more blood pressure (BP) value associated with the patient; Vule further discloses that the one or more determined feature values are input to the trained machine learning model (FIG. 10, component 1006; paragraph [0116]) which analyzes the feature data and outputs an estimate of blood pressure associated with the patient (paragraph [0116]).
d) outputting the one or more BP value determined for the patient based on the determined one or more features. Vule discloses that the machine learning model analyzes the determined feature values and outputs an estimate of blood pressure associated with the patient (paragraph [0116]; FIG. 10).
While Vule discusses that PPG-related signals (FIG. 9, component 902) are used to train the deep learning model (FIG. 9, component 904) that then generates temporal feature data and morphological data to train the machine learning model for determining a BP value of a user (FIG. 9, component 910; paragraph [0097], Vule does not specifically disclose that the training feature data is from a plurality of prior patients or that the plurality of prior patients comprises hemodynamically compromised patients.
However, Li teaches a system and method comprising the use of a machine learning model for the estimation of systolic and diastolic blood pressure values using ECG and PPG signals, where the model is trained using features (Table 1; FIG. 6) extracted from datasets that comprise a compilation of prior patient data known as multivariate intelligent monitoring in intensive care II (MIMIC II) that includes over 25,000 instances that comprise a patient’s record and measured biomedical signals, such as PPG signals, ECG signals, and arterial blood pressure measurements, for each instance (Section: “3. Methodology: 3.8. Overall Framework”; Section: “4. Experiment: 4.1 Data Set”, paragraph 1). Li teaches that the data from MIMIC II are processed and then 80% of the processed data sets are used as a training set for the machine learning model (Section: “4. Experiment: 4.1 Data Set”, paragraph 5).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date to have modified the system and method disclosed by Vule, which includes a trained model for determining a BP value of a user based on PPG signal data with the teaching of Li that a BP prediction model can be trained using training temporal and morphology feature data from a plurality of patients because it would be considered combining prior art elements according to known methods to yield the predictable result of training a machine learning model for determining a BP value of a user.
Li discusses the use of the MIMIC II dataset as the basis for training the BP estimation model, but does not provide further details of the MIMIC II data set including what information about the plurality of prior patients is represented in the MIMIC II dataset. However, as evidenced by Clifford (Clifford, G., et al., User Guide and Documentation for the MIMIC II Data Base, 24 Feb 2012), which discloses the contents of the MIMIC II data set, the MIMIC II dataset includes data from prior patients that are hemodynamically compromised (Pages 64 – 65, Section: “6 Appendix: 6.4 Frequency of all ICD-9 codes for adult ICU-related hospital admissions”) including prior patients with hypertensive disease, heart failure, artery and vein disorders, diseases of pulmonary circulation, and heart disease.
In regard to claim 2, Vule discloses the invention of claim 1, further comprising a user interface. Vule discloses that the system includes a user interface (FIG. 8, component 850) that can include a display comprising a graphical user interface (paragraph [0091]).
In regard to claim 3, Vule discloses the invention of claim 1, wherein at least one of the at least one sensor, the at least one processor, and the at least one computer-readable storage medium is wearable. Vule specifically discloses that the system includes a wearable device (FIG. 1) that includes the at least one sensor, such as the PPG sensor, processor, and memory storage containing executable instructions for performing the method of determining a blood pressure estimate (FIG. 9) using feature data collected from the wearable sensors (paragraph [0120]).
In regard to claim 4, Vule discloses the invention of claim 1, wherein the at least one sensor comprises a photoplethysmography (PPG) sensor or an electrocardiogram (ECG) sensor. Vule discloses that the wearable device comprises a plurality of sensors (FIG. 8, component 870) including PPG sensors (FIG. 8, see “Optical Sensors”; paragraph [0098]) and ECG sensors (FIG. 8, see “ECG Electrode(s)).
In regard to claim 5, Vule discloses the invention of claim 1, wherein each of the one or more BP value comprises one or more of a systolic BP, a diastolic BP, and/or a mean arterial pressure (MAP). Vule discloses that the one or more determined BP values include systolic BP and diastolic BP (paragraph [0116]).
In regard to claim 7, Vule discloses the invention of claim 1, wherein the one or more trained models comprises one or more of: a lasso model, a random forest model, a support vector machine model, an artificial neural network model, a long short term memory model, a RESNET deep learning model, and/or a combination or ensemble thereof. Vule discloses the use of two trained models, a first model used to obtain features related to BP of a user (FIG. 13, component 1304) that comprises an artificial neural network or more specifically a convolution neural network (paragraphs [0095] & [0128]), and a second model used to estimate the BP of a user using feature values that comprises a random forest model or a support vector machine model (paragraph [0035]).
In regard to claim 10, Vule discloses the invention of claim 1, wherein the one or more morphology features are obtained from at least one PPG waveform, wherein the at least one PPG waveform is collected noninvasively over a radial artery. Vule discloses that the PPG waveform is collected using sensors (FIG. 3, component 150, see “Radial Artery Center”) positioned at the center of the user’s radial artery (paragraph [0051]) which allows the system to detect, collect, and analyze blood flow through the radial artery using reflective or transmissive systems such as PPG sensors (paragraph [0060]).
In regard to claim 11, Vule discloses the invention of claim 1 wherein the one or more features further comprises biometric data, wherein the biometric data comprises one or more of: a pre- existing condition, an age, a weight, a height, a waist size, a body mass index (BMI), a sex, and/or a combination thereof. Vule discloses that feature information comprising personal-related information such as sex, age, weight, race, and any other medically meaningful distinction or information such as a pre-existing condition can be input to the model to determine a BP value of a user (paragraph [0114]; FIG. 9, component 916).
In regard to claim 15, Vule discloses the invention of claim 1, wherein the one or more temporal features or the one or more morphology features are extracted from a first PPG sensor and a first ECG sensor (FIG. 8, see “Optical Sensor(s)” and “ECG Electrode(s); paragraph [0026]).
In regard to claim 17, Vule as modified discloses the invention of claim 1. Li teaches the use of the MIMIC II dataset for training a model that output BP estimations (Abstract) which includes data collected from the intensive care unit (ICU) of Boston’s Beth Israel Deaconess Medical Center (Clifford, Page 3, Section: “1 Introduction: 1.3 Overview of data collection) as evidenced by Clifford. One of ordinary skill in the art would recognize that each of the prior patients represented in the MIMIC II dataset would have a pre-existing condition, such as those outlined in Pages 64 – 65, Section: “6 Appendix: 6.4 Frequency of all ICD-9 codes for adult ICU-related hospital admissions” of Clifford, to have been admitted to the ICU for medical treatment.
In regard to claims 23, 41, & 59, Vule discloses a method for estimating blood pressure, the method comprising:
a) measuring, using the at least one sensor, at least one waveform related to blood pressure from a patient. Vule discloses that sensors (FIG. 8, component 870) are used to measure a biological signal, such as a PPG waveforms (FIG. 11, paragraph [0118]) and ECG waveforms (paragraph [0030]).
b) receiving, from the at least one sensor, the at least one waveform and determining, from the at least one waveform, one or more features comprising at least one or more temporal features and or one or more morphology features; Vule discloses that a wrist worn PPG sensor (paragraph [0061]; FIG. 1, component 165) is used to collect PPG waveform data (paragraph [0066]) and extract a number of features from the collected blood flow pulse signal, such as shape of the pulse, a maximum amplitude, a minimum amplitude, a maximum derivative, a time difference between main and secondary peaks, and integral through the entire extraction time (paragraph [0067]; FIGs. 6A - 6C).
c) determining, based on the determined feature set of one or more features from a patient, one or more blood pressure (BP) value associated with the patient, the determining comprising analyzing the one or more features using one or more trained models, wherein each of the one or more trained models are trained using at least training temporal feature data and training morphology data; Vule discloses a blood pressure assessment method with a plurality of sensors (FIG. 8, component 870) used to measure a biological signal, such as a PPG waveforms (FIG. 11, paragraph [0118]) and ECG waveforms (paragraph [0030]), which is used to isolate one or more morphology features related to blood pressure and estimate a physiological property, such as blood pressure, using a trained machine learning model (paragraphs [0007] - [0008]).
d) outputting the one or more BP value determined for the patient based on the received feature set. Vule further discloses that the method includes outputting the systolic and diastolic blood pressure values of a user (FIG. 10).
Vule further discloses the at least one non-transitory storage medium having encoded thereon executable instructions of claim 41 that, when executed by at least one processor, cause the at least one processor to carry out the method described above (paragraph [0086]) and a system comprising at least one processor (paragraph [0087], FIG. 8, component 805) and at least one computer-readable storage medium of claim 59 to carry out the method described above (paragraph [0086]).
While Vule discusses that PPG-related signals (FIG. 9, component 902) are used to train the deep learning model (FIG. 9, component 904) that then generates temporal feature data and morphological data to train the machine learning model for determining a BP value of a user (FIG. 9, component 910; paragraph [0097], Vule does not specifically disclose that the training feature data is from a plurality of prior patients or that the plurality of prior patients comprises hemodynamically compromised patients.
However, Li teaches a system and method comprising the use of a machine learning model for the estimation of systolic and diastolic blood pressure values using ECG and PPG signals, where the model is trained using features (Table 1; FIG. 6) extracted from datasets that comprise a compilation of prior patient data known as multivariate intelligent monitoring in intensive care II (MIMIC II) that includes over 25,000 instances that comprise a patient’s record and measured biomedical signals, such as PPG signals, ECG signals, and arterial blood pressure measurements, for each instance (Section: “3. Methodology: 3.8. Overall Framework”; Section: “4. Experiment: 4.1 Data Set”, paragraph 1). Li teaches that the data from MIMIC II are processed and then 80% of the processed data sets are used as a training set for the machine learning model (Section: “4. Experiment: 4.1 Data Set”, paragraph 5).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date to have modified the system and method disclosed by Vule, which includes a trained model for determining a BP value of a user based on PPG signal data with the teaching of Li that a BP prediction model can be trained using training temporal and morphology feature data from a plurality of patients because it would be considered combining prior art elements according to known methods to yield the predictable result of training a machine learning model for determining a BP value of a user.
Li discusses the use of the MIMIC II dataset as the basis for training the BP estimation model, but does not provide further details of the MIMIC II data set including what information about the plurality of prior patients is represented in the MIMIC II dataset. However, as evidenced by Clifford, which discloses the contents of the MIMIC II data set, the MIMIC II dataset includes data from prior patients that are hemodynamically compromised (Pages 64 – 65, Section: “6 Appendix: 6.4 Frequency of all ICD-9 codes for adult ICU-related hospital admissions”) including prior patients with hypertensive disease, heart failure, artery and vein disorders, diseases of pulmonary circulation, and heart disease.
In regard to claim 82, Vule as modified discloses the invention of claim 1. Vule further discloses that one or more features including systolic peak or peak volume can be extracted from PPG signals (paragraphs [0117] – [0118]; FIG. 11, component “S”) and Li further teaches that training morphology data features are extracted from the plurality of prior patients represented in the MIMIC II data set (Section: “3. Methodology: 3.8. Overall Framework”; Section: “4. Experiment: 4.1 Data Set”, paragraph 1).
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) as applied to claim 1 above, and further in view of Centen (US 20190046069 A1 – Previously Cited).
In regard to claim 6, Vule discloses the invention of claim 1. While Vule discloses that a variety of morphological and temporal features can be utilized to determine the BP value of a user, including features extracted from single pulses and series of pulses such as statistical averages of various features across a series, information regarding the morphological shape of each pulse, the average and standard deviation of morphology of a series of pulses, temporal features such as the timing of various features within single pulses, the duration of a single pulse, as well as the average and standard deviation of the timing of a feature or duration of pulses within a series of pulses, and the timing of morphological features across a series of pulses (paragraph [0071]), they do not specify that the determined one or more features is a preprocessed feature set, the preprocessed feature set comprising the one or more features configured as a function of one or more of: a pre-ejection period, a square of pulse transit time, a PPG intensity ratio, and/or a Womersly number.
However, Centen teaches a system that extracts a pre-ejection period (PEP) of a user from ECG data (paragraph [0115]; FIG. 7D) and inputs the generated cardiovascular health parameter or feature values into a cardiovascular transformation model (FIG. 8, component 841 and “Cardiovascular Transformation Model”) to generate a clinical parameter (FIG. 8, component 842) including a BP value of a user (paragraph [0131]; FIG. 7A, component 729).
It would have been obvious to one of ordinary skill in art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule with the teaching of Centen that a system can extract feature data from ECG signals including pre-ejection period and input the feature data to a model to generate a BP value of a user because Vule already discusses that a plurality of morphological and temporal features can be input to a trained model to generate a BP value of a user such that the modification of the system disclosed by Vule with the teaching of Centen would be considered combining prior art elements according to known methods to yield the predictable result of generating a BP value of a user based on feature data.
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) as applied to claim 1 above, and further in view of Goldner (US 20220361823 A1 – Previously Cited).
In regard to claim 8, Vule discloses the invention of claim 1. While Vule discloses one or more trained models that are used to output a BP value associated with a user and further discusses improving model performance by preventing overfitting results to training data (paragraphs [0033] - [0035]), they do not specify that analyzing comprises tuning hyperparameters for each of the one or more trained models.
However, Goldner teaches that performance of a machine learning model used to predict a BP value of a user (FIG. 3A, component 312) can adjust or fine-tune parameters of the model, including modifying hyperparameters, in order to improve the performance or accuracy of the model (paragraph [0085]; FIG. 3A, component 314).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to have modified the machine learning model for determining a BP value of a user with the teaching that a machine learning model for determining a BP value of a user can be adjusted or tuned by modifying the hyperparameters of the model because adjusting the hyperparameters of the model improves the performance of the model (Goldner, paragraph [0085]).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) as applied to claim 1 above, and further in view of de Groot (US 20240298903 A1 – Previously Cited).
In regard to claim 9, Vule discloses the invention of claim 1. While Vule discloses that a variety of morphological and temporal features can be utilized to determine the BP value of a user, including features extracted from single pulses and series of pulses, such as statistical averages of various features across a series, information regarding the morphological shape of each pulse, the average and standard deviation of morphology of a series of pulses, temporal features such as the timing of various features within single pulses, the duration of a single pulse, as well as the average and standard deviation of the timing of a feature or duration of pulses within a series of pulses, and the timing of morphological features across a series of pulses (paragraph [0071]), they do not specify that the one or more morphology features comprise: a BP cycle time, an ejection time, an artery fill time, an artery emptying time, a peak volume, a systolic volume, a systolic volume differential, a diastolic volume, a diastolic volume differential, or a combination thereof.
However, de Groot teaches that left ventricular ejection time can be used as a feature value that is input to a machine learning model for determining a BP value of a user (paragraph [0045]; FIG. 4).
It would have been obvious to one of obvious skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule with the teaching of de Groot that left ventricular ejection time can be used as a feature value for determining a BP value of a user because Vule already discusses the use of different morphological and temporal features including the timing of morphological features such that modifying the system disclosed by Vule with the teaching of de Groot would be considered combining prior art elements according to known methods to yield the predictable result of predicting a BP value of a user from feature data.
Claims 12 & 14 are rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) as applied to claim 1 above, and further in view of Pantelopoulos (US 20170209053 A1 – Previously Cited).
In regard to claim 12, Vule discloses the invention of claim 1. While Vule discloses that a variety of morphological and temporal features can be utilized to determine the BP value of a user, including features extracted from single pulses and series of pulses such as statistical averages of various features across a series, information regarding the morphological shape of each pulse, the average and standard deviation of morphology of a series of pulses, temporal features such as the timing of various features within single pulses, the duration of a single pulse, as well as the average and standard deviation of the timing of a feature or duration of pulses within a series of pulses, and the timing of morphological features across a series of pulses (paragraph [0071]), they do not specify that the one or more temporal features comprises one or more of: a pulse arrival time (PAT), a pulse transit time (PTT), a pulse rate, and/or a combination thereof.
However, Pantelopoulos teaches that pulse transit time and pulse arrival time (TABLE 00001) can be isolated from sensor signals, such as PPG signals (FIG. 1A) or ECG signals (paragraph [0174]) and input as feature values to a model to determine a BP value of a user (paragraph [0009]).
It would have been obvious to one of obvious skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule with the teaching of Pantelopoulos pulse transit time and pulse arrival time can be used as feature values for determining a BP value of a user because Vule already discusses the use of different morphological and temporal features including the timing of morphological features such that modifying the system disclosed by Vule with the teaching of Pantelopoulos would be considered combining prior art elements according to known methods to yield the predictable result of predicting a BP value of a user from feature data.
In regard to claim 14, Vule as modified discloses the invention of claim 12, wherein the PTT is based on a difference between at least two PPG waveforms. Pantelopoulos teaches that PTT is determined by calculating the difference between a proximal PPG waveform and a distal PPG waveform (paragraphs [0083] - [0084]; FIG. 1A).
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) and further in view of Pantelopoulos (US 20170209053 A1 – Previously Cited) as applied to claim 12 above, and further in view of Wang (US 20230320599 A1 – Previously Cited).
In regard to claim 13, Vule as modified discloses the invention of claim 12. While Vule discloses that different morphological and temporal features including the timing of morphological features can be used as inputs to a model to determine a BP value and Pantelopoulos further teaches that PPG data and ECG data can be used to determine pulse arrival time (PAT), they do not explicitly discuss that the PAT is based on a time difference between a peak of an ECG-R wave and a peak of a PPG waveform.
However, Wang teaches that PAT is estimated from the time delay or difference between an ECG R peak to PPG main wave feature points (FIG. 3, components 126 & 128; FIG. 4; paragraph [0038]).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule as modified by Pantelopoulos with the teaching of Wang that PAT is determined from the time difference of an ECG R peak to PPG main wave feature points because Pantelopoulos already teaches that PPG data and ECG data can be used to determine pulse arrival time (PAT). The modification of Vule as modified with Pantelopoulos with the teaching of Wang would be considered combining prior art elements according to known methods to yield the predictable result of determining PAT from ECG and PPG data.
Claims 18 & 20 - 22 are rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) as applied to claim 1 above, and further in view of Varadan (US 20230082362 A1 – Previously Cited).
In regard to claim 18, Vule discloses the invention of claim 1. While Vule discloses the determining of one or more features from at least one waveform and further determining one or more BP values using a trained model, Vule does not disclose that that the determining comprises selecting the one or more trained models from a plurality of models, wherein the selecting is based on evaluating a performance of the one or more trained models as compared to others of the plurality of models.
However, Varadan teaches a method for generating and evaluating machine learning models for BP estimation where a plurality of models are trained and compared to one another based on the amount of error associated with the estimated output for model selection (FIG. 1, component 750; paragraphs [0137] - [0138]).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule with the teaching of Varadan that multiple trained models are prepared and the performance of each model is compared to one another because doing so allows for the selection of the best performing model for the task of generating a BP value of a user (Varadan, paragraph [0138]).
In regard to claim 20, Vule as modified discloses the invention of claim 18, wherein the evaluating is performed independently for each of the one or more BP value. Varadan teaches that as each set of input data is presented, the model determines an output of estimated BP, which is then evaluated or compared to a ground truth value to generate an error value associated with the estimated BP value (paragraph [0137]).
In regard to claim 21, Vule discloses the invention of claim 1, wherein the analyzing comprises an ensemble of two or more of the one or more trained models. Vule discloses the use of an ensemble of two or more trained models including a deep learning model (FIG. 10, component 1004) and a machine learning model (FIG. 10, component 1006) which work in sequence to output the systolic and diastolic blood pressure values of the user (FIG. 10). While Vule discloses the use of an ensemble of two or more trained models, Vule does not disclose that that the analyzing comprises selecting an ensemble comprising two or more of the one or more trained models, wherein the selecting is based on evaluating a performance of the two or more of the one or more trained models as compared to others of the one or more trained models.
However, Varadan teaches a method for generating and evaluating machine learning models for BP estimation where a plurality of models are trained and compared to one another based on the amount of error associated with the estimated output for model selection (FIG. 1, component 750; paragraphs [0137] - [0138]).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule, which includes an ensemble of trained models, with the teaching of Varadan that multiple trained models are prepared and the performance of each model is compared to one another because doing so allows for the selection of the best performing model for the task of generating a BP value of a user (Varadan, paragraph [0138]).
In regard to claim 22, Vule as modified discloses the invention of claim 21, wherein the ensemble comprises the two or more of the one or more trained models being in parallel or serial. Vule discloses the use of an ensemble of two or more trained models including a deep learning model (FIG. 10, component 1004) and a machine learning model (FIG. 10, component 1006) which work in sequence or serial to output the systolic and diastolic blood pressure values of the user (FIG. 10).
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Vule (US 20220183569 A1 – Previously Cited) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) and further in view of Varadan (US 20230082362 A1) as applied to claim 18 above, and further in view of Wu (Wu H, Ji Z, Li M. Non-Invasive Continuous Blood-Pressure Monitoring Models Based on Photoplethysmography and Electrocardiography. Sensors (Basel). 2019;19(24):5543. Published 2019 Dec 15. doi:10.3390/s19245543 – Previously Cited).
In regard to claim 19, Vule as modified discloses the invention of claim 18. While Varadan teaches that models may be compared to one another based on a variety of parameters, including mean squared error, mean absolute error, or r squared values, they do not specify that the evaluating is based on one or more of an average error bias and/or a standard deviation of each of the plurality of models.
However, Wu teaches a method for generating and evaluating machine learning models for non-invasive continuous BP monitoring based on PPG and EEG data. The models, including a multiple population genetic algorithm (MPGA) Mean Impact Value (MIV) BP model and partial least squares regression (PLSR) model, were compared to one another (FIG. 8; Table 9) based on standard deviation (STD), mean absolute deviation (MAD), and root mean square error (RMSE).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to have modified the system disclosed by Vule as modified with the teaching of Wu that models can be compared to one another using standard deviation values because Varadan already teaches that trained machine learning models can be compared to one another based on a variety of performance metrics, such that it would be considered combining prior art elements, in this case the performance parameters that are used to assess and compare different machine learning models, to yield the predictable result of comparing the performance of different models for generating a BP value and selecting the best model.
Response to Arguments
Applicant’s arguments, see Remarks, filed 07/20/2026, with respect to the 112(b) rejections of claims 18 - 22 have been fully considered and are persuasive. The 112(b) rejections of claims 18 - 22 have been withdrawn.
Applicant’s arguments, see Remarks, filed 07/20/2026, with respect to the 101 rejections of claims 1 – 6, 8 – 23, 41, and 59 have been fully considered and are persuasive. The 101 rejections of claims 1 – 6, 8 – 23, 41, and 59 have been withdrawn.
Applicant’s arguments, see Remarks, filed 07/20/2026, with respect to the rejections of claims 1 – 5, 7, 10, 11, and 15 under 35 U.S.C. 102 and claims 6, 8, 9, 12, 13, 14, 17 – 19, 23, 41, and 59 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection of independent claims 1, 23, 41, and 59 under 35 U.S.C. 103 is made in view of Vule (US 20220183569 A1) in view of Li (Li Y, et al., Real-Time Cuffless Continuous Blood Pressure Estimation Using Deep Learning Model. Sensors (Basel). 2020;20(19):5606. 2020 Sep 30) that further apply to the dependent claims.
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 SIENNA CHRISTINE PYLE whose telephone number is (703)756-5798. The examiner can normally be reached 8 am - 5:30 pm M - T; Off first Fridays; 8 am - 4 pm second Fridays.
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, Charles Marmor, II can be reached at (571) 272-4730. 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.
/ERIC F WINAKUR/Primary Examiner, Art Unit 3791
/S.C.P./Examiner, Art Unit 3791
