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
This Office Action is responsive to the amendment filed on 02 Jun 2026. As directed by the amendment: no claims have been amended, claims 1-14 have been canceled, and claims 21-34 have been added. Thus, claims 15-34 are presently pending in this application.
Drawings
The drawings are objected to as failing to comply with 37 CFR 1.84(p)(5) because they include the following reference character(s) not mentioned in the description: 310 in Fig. 3. Corrected drawing sheets in compliance with 37 CFR 1.121(d), or amendment to the specification to add the reference character(s) in the description in compliance with 37 CFR 1.121(b) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
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 15-34 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?
Claims 15-31 recite a machine (system), claims 32-33 recite a process (method), and claim 34 recites a manufacture (non-transitory machine-readable medium), which are within the four statutory categories. Therefore, claims 15-34 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 15-34 are directed to an abstract idea.
Claim 32 is directed to a method comprising: determining an acoustic signal snippet that corresponds to at least a portion of acoustic signal information associated with a user; applying one or more signal processing algorithms to the acoustic signal snippet to determine one or more acoustic features; producing a representation of the one or more acoustic features using a trained machine learning model; and classifying the representation using the trained machine learning model to predict one or more physiological states of the user.
Claim 15 is directed to a system comprising a processor and a memory device containing instructions, which when executed by the processor, cause the processor to perform the operations recited in claim 32.
Claim 34 is directed to a non-transitory machine-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to perform the operations recited in claim 32.
The limitations of determining an acoustic signal snippet, applying one or more signal processing algorithms, producing a representation using a machine learning model, and classifying the representation using the 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, “determining an acoustic signal snippet that corresponds to at least a portion of acoustic signal information associated with a user” encompasses nothing more than a human actor mentally evaluating acoustic signal data according to certain rules in order to select pieces of data.
The limitation “producing a representation of the one or more acoustic features using a trained machine learning model” encompasses nothing more than a human actor drawing out acoustic features on a piece of paper, and/or mentally evaluating the acoustic features to determine a representation.
Therefore, claims 15, 32, and 34 recite an abstract idea.
Claims 16-31 depend on claim 15, and claim 33 depends on claim 32. These dependent claims only recite additional features of the analysis described in claims 15 and 32, which may also be performed by a human actor mentally and using a pen and paper. For example, claim 17 recites “segmenting the acoustic signal snippet into a plurality of segments, wherein each of the plurality of segments corresponds to a physiological state zone indicating a specific physiological event of the user”, which encompasses nothing more than a human actor evaluating the acoustic signal snippet to determine which segments of the snippet correspond to a physiological state zone indicating a specific physiological event of the user.
Therefore, claims 15-34 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 15 only recites the additional limitations “a processor” and “a memory device”. Claim 34 recites “a non-transitory machine-readable medium” and “one or more processors”. These additional elements are recited at a high level of generality (i.e. most generic computers would be known to have these components). Paragraphs [0099]-[0100], [0104]-[0105], and [0107] describe the memory and processors at a high level of generality. These generic processor and memory 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 claims 15 and 34 do not integrate the judicial exception into a practical application.
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 16-31 and 33 only recite other limitations of the mental processes recited in claims 15 and 32, 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. Claim 15 only recites the additional limitations “a processor” and “a memory device”. Claim 34 recites “a non-transitory machine-readable medium” and “one or more processors”. These additional elements are recited at a high level of generality (i.e. most generic computers would be known to have these components). Paragraphs [0099]-[0100], [0104]-[0105], and [0107] describe the processor and non-transitory computer-readable medium 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.
Therefore, claims 15-34 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 § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 15-19, 25-26, and 28-34 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Zhou et al. (US 20230329646 A1), hereinafter Zhou.
Regarding claim 15, Zhou discloses a system (Fig. 1, paragraph [0043], information processing system 100) comprising:
a processor (Fig. 1, paragraph [0070], processor 120); and
a memory device containing instructions (Fig. 1, paragraph [0073], memory 122), which when executed by the processor, cause the processor to perform operations comprising:
determining an acoustic signal snippet that corresponds to at least a portion of acoustic signal information associated with a user (paragraph [0057], "data from one or more of the acoustic signal sources 105"; paragraph [0085], "In the sound capture component 202, sound produced from the body is captured, for example, by a transducer or other type of sensor, such as a sensor associated with an intelligent stethoscope or other patient monitoring device"; paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
applying one or more signal processing algorithms to the acoustic signal snippet to determine one or more acoustic features (paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
producing a representation of the one or more acoustic features using a trained machine learning model (paragraph [0090], "2D-CNN 210 is trained to discern sounds as a human would, since the established clinical protocol of what to do depending on the sound heard is formulated around the sounds that doctors or other medical professionals can perceive"; paragraph [0099], "The 2D-CNN 210 is forced to learn multiple independent representations of the same data"); and
classifying the representation using the trained machine learning model to predict one or more physiological states of the user (paragraph [0046], "process the image representation in at least one neural network of the biomedical acoustics classifier 110 to generate a classification for the acoustic signal"; paragraph [0043], "biomedical acoustics classifier adapted to classify acoustic signals in one or more designated physiological contexts, such as patient diagnosis"; paragraph [0050], "The generated classification can comprise, for example, an indicator of a particular detected physiological condition of the given individual").
Regarding claim 16, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the operations further comprise applying the one or more signal processing algorithms to the acoustic signal snippet to obtain the one or more acoustic features having at least one of a Mel spectrogram parameter, a Mel frequency cepstral coefficient parameter (paragraph [0110], "layered architecture 700 receives as an input a Mel-scaled, dB-scaled spectrogram 701"), or a power spectral density parameter (paragraph [0113], "Referring initially to FIG. 8A, a biomedical acoustics classifier 800 utilizing a spectrogram and a power spectrum is shown").
Regarding claim 17, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the operations further comprise segmenting the acoustic signal snippet into a plurality of segments, wherein each of the plurality of segments corresponds to a physiological state zone indicating a specific physiological event of the user (paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds").
Regarding claim 18, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the operations further comprise applying a combination of the one or more acoustic features into one or more neural network branches of the trained machine learning model, wherein the trained machine learning model is produced based on the combination of the one or more acoustic features (paragraph [0117], "The example parallel neural network structure of the FIG. 8A embodiment allows acoustic signal information encoded in the time domain and frequency domain to be analyzed simultaneously and the learned features to be merged together in determining the final classification"; paragraph [0127], "some embodiments are configured to utilize a merged CNN and RNN architecture. Such an embodiment leverages the long-term dependencies that are present in some types of acoustic signals, such as heart sound signals").
Regarding claim 19, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the operations further comprise applying different pairings of the one or more acoustic features into respective neural network branches of the trained machine learning model, wherein the trained machine learning model is produced based on the different pairings of the one or more acoustic features (paragraph [0052], "In some embodiments, generating the image representation illustratively comprises generating at least one spectrogram, with each such spectrogram representing frequency, time and amplitude in respective dimensions thereof. Other types of image representations can be used in other embodiments. For example, some embodiments disclosed herein utilize recurrence plots in addition to or in place of spectrograms. As further examples, additional image representations can be generated using image data augmentation techniques including but not limited to geometric transformations, color-space transformations, masking, kernel filters, and/or other techniques not available at the audio signal level, to expand a training set for processing by a neural network. Such data augmentation techniques can be applied to spectrograms, recurrence plots, or other types of image representations"; paragraph [0107], "Both the real and artificially created spectrograms are used to train the 2D-CNN 210 for classification"; paragraph [0113], "Referring initially to FIG. 8A, a biomedical acoustics classifier 800 utilizing a spectrogram and a power spectrum is shown"; paragraph [0118], "Referring now to FIG. 8B, a biomedical acoustics classifier 801 utilizing both a spectrogram and a recurrence plot is shown").
Regarding claim 25, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the trained machine learning model has been produced by a training neural network using the one or more acoustic features as input to predict the one or more physiological states of the user (paragraph [0103], "another synthetic data generation method that can be applied in illustrative embodiments disclosed herein involves using the spectrogram images to train a Generative Adversarial Network (GAN), and then using the trained GAN to generate new, synthetic spectrograms to train the CNN").
Regarding claim 26, Zhou discloses the system of claim 25, as explained above. Zhou further discloses that the neural network has been trained using supervised learning (paragraph [0301], "Illustrative embodiments herein provide deep learning models for automated AVF stenosis screening based on the sound of AVF blood flow using supervised learning with data validated by ultrasound"; paragraph [0303], "The disclosed models are trained using supervised learning with labels validated from concurrent duplex ultrasound").
Regarding claim 28, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that determining the acoustic signal snippet comprises applying a sliding window having a predefined window length and a predefined stride length to an acoustic signal (paragraph [0087], "windowing is accomplished using a Hann window of size 512 and hop length of 256"; paragraph [0256], "The convolution operator is a sliding window or kernel that performs local aggregations across neighboring pixels").
Regarding claim 29, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the trained machine learning model comprises a convolutional neural network (paragraph [0084], two-dimensional (2D) convolutional neural network (CNN) 210) including convolution layers (Fig. 6, paragraph [0108], convolutional layers 602, 606), max-pooling layers (Fig. 6, paragraph [0108], max pooling layers 604, 608), a flattening layer (Figs. 8A-8B, paragraphs [0114], [0116], [0118], flattened layer 813), fully connected layers (Fig. 6, paragraph [0108], fully connected layer 610; Figs. 8A-8B, paragraphs [0114], [0116], [0118], fully connected layer 814) with dropout (paragraph [0126], "The output is then passed through a dropout layer"), and a softmax layer (Fig. 6, paragraph [0108], softmax layer 612).
Regarding claim 30, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the trained machine learning model has been trained using acoustic signal data comprising phonocardiogram data associated with one or more users as training data to predict the one or more physiological states (paragraph [0183], "An objective of these experiments was to identify optimal forms of data augmentation for illustrative embodiments in the binary classification of PCG signals using their spectral image representation").
Regarding claim 31, Zhou discloses the system of claim 15, as explained above. Zhou further discloses that the acoustic signal snippet corresponds to a fixed-duration segment of an original acoustic signal (paragraph [0087], "windowing is accomplished using a Hann window of size 512"; paragraph [0269], "the audio signals are windowed using a Hann window of size 512").
Regarding claim 32, Zhou discloses a method comprising:
determining an acoustic signal snippet that corresponds to at least a portion of acoustic signal information associated with a user (paragraph [0057], "data from one or more of the acoustic signal sources 105"; paragraph [0085], "In the sound capture component 202, sound produced from the body is captured, for example, by a transducer or other type of sensor, such as a sensor associated with an intelligent stethoscope or other patient monitoring device"; paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
applying one or more signal processing algorithms to the acoustic signal snippet to determine one or more acoustic features (paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
producing a representation of the one or more acoustic features using a trained machine learning model (paragraph [0090], "2D-CNN 210 is trained to discern sounds as a human would, since the established clinical protocol of what to do depending on the sound heard is formulated around the sounds that doctors or other medical professionals can perceive"; paragraph [0099], "The 2D-CNN 210 is forced to learn multiple independent representations of the same data"); and
classifying the representation using the trained machine learning model to predict one or more physiological states of the user (paragraph [0046], "process the image representation in at least one neural network of the biomedical acoustics classifier 110 to generate a classification for the acoustic signal"; paragraph [0043], "biomedical acoustics classifier adapted to classify acoustic signals in one or more designated physiological contexts, such as patient diagnosis"; paragraph [0050], "The generated classification can comprise, for example, an indicator of a particular detected physiological condition of the given individual").
Regarding claim 33, Zhou discloses the method of claim 32, as explained above. Zhou further discloses that the trained machine learning model has been trained using acoustic signal data comprising phonocardiogram data associated with one or more users as training data to predict the one or more physiological states (paragraph [0183], "An objective of these experiments was to identify optimal forms of data augmentation for illustrative embodiments in the binary classification of PCG signals using their spectral image representation").
Regarding claim 34, Zhou discloses a non-transitory machine-readable medium comprising instructions (Fig. 1, paragraph [0073], memory 122) that, when executed by one or more processors (Fig. 1, paragraph [0070], processor 120), cause the one or more processors to perform operations comprising:
determining an acoustic signal snippet that corresponds to at least a portion of acoustic signal information associated with a user (paragraph [0057], "data from one or more of the acoustic signal sources 105"; paragraph [0085], "In the sound capture component 202, sound produced from the body is captured, for example, by a transducer or other type of sensor, such as a sensor associated with an intelligent stethoscope or other patient monitoring device"; paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
applying one or more signal processing algorithms to the acoustic signal snippet to determine one or more acoustic features (paragraph [0032], "Segmentation is the process of identifying the positions and boundaries of S1, S2, systole, and diastole in the cardiac cycle, where S1 and S2 denote respective first and second heart sounds"; paragraph [0087], "In the spectrogram generation component 206, the sound signal is segmented into overlapping windows");
producing a representation of the one or more acoustic features using a trained machine learning model (paragraph [0090], "2D-CNN 210 is trained to discern sounds as a human would, since the established clinical protocol of what to do depending on the sound heard is formulated around the sounds that doctors or other medical professionals can perceive"; paragraph [0099], "The 2D-CNN 210 is forced to learn multiple independent representations of the same data"); and
classifying the representation using the trained machine learning model to predict one or more physiological states of the user (paragraph [0046], "process the image representation in at least one neural network of the biomedical acoustics classifier 110 to generate a classification for the acoustic signal"; paragraph [0043], "biomedical acoustics classifier adapted to classify acoustic signals in one or more designated physiological contexts, such as patient diagnosis"; paragraph [0050], "The generated classification can comprise, for example, an indicator of a particular detected physiological condition of the given individual").
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 20-21 are rejected under 35 U.S.C. 103 as being unpatentable over Zhou et al. (US 20230329646 A1), hereinafter Zhou, in view of Pahlevan et al. (US 20230420132 A1), hereinafter Pahlevan.
Regarding claim 20, Zhou discloses the system of claim 15, as explained above. Zhou does not explicitly disclose that the operations further comprise: determining a metric indicating a comparison between a physiological state prediction and a target physiological state to evaluate the trained machine learning model; determining whether the metric exceeds an error threshold; and updating the trained machine learning model when the metric exceeds the error threshold.
However, Pahlevan teaches non-invasive techniques for determining whether a patient has experienced heart failure using a machine learning model (Abstract) based on phonocardiogram data (paragraphs [0049], [0060]), wherein the operations comprise:
determining a metric indicating a comparison between a physiological state prediction and a target physiological state to evaluate the trained machine learning model (paragraph [0103], accuracy; paragraph [0136], "evaluation metrics included sensitivity, specificity, accuracy, and the area under the curve (AUC) defined by receiver-operating characteristic (ROC) analysis");
determining whether the metric exceeds an error threshold (paragraph [0103], "model training/testing module 914 may execute one or more optimization techniques, such as gradient descent, to minimize an error of each machine learning model until the model's accuracy is determined to reach some threshold criteria (e.g., 75% or greater accuracy, 80% or greater accuracy, 90% or greater accuracy, or other accuracy values)"); and
updating the trained machine learning model when the metric exceeds the error threshold (paragraph [0103], "the updating/retraining may occur in response to determining that a model fails to produce accurate results (e.g., its accuracy falls below a threshold accuracy)").
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 Zhou with the teachings of Pahlevan so that the operations further comprise: determining a metric indicating a comparison between a physiological state prediction and a target physiological state to evaluate the trained machine learning model; determining whether the metric exceeds an error threshold; and updating the trained machine learning model when the metric exceeds the error threshold, because doing so minimizes the error of the machine learning model (Pahlevan, paragraph [0103]), ensuring that the model performs accurately Pahlevan, paragraph [0142]).
Regarding claim 21, the system of claim 20 is obvious over Zhou and Pahlevan, as explained above. Zhou further discloses using a cross-entropy loss as a criterion for optimization (paragraph [0277], "The model is trained using Adam optimizer at a learning rate of 1×10.sup.-3 over the binary cross-entropy loss function in the case of binary classification and over the categorical cross-entropy loss function in the case of multiclass classification").
Pahlevan further teaches that the trained machine learning model has been updated by adjusting model weight parameters based on the metric as an auxiliary criterion (paragraph [0159], "If the accuracy score of the model is greater than or equal to an accuracy score threshold (e.g., an 80% or greater accuracy, a 90% or greater accuracy, a 95% or greater accuracy, etc.), then process 2400 may proceed to operation 2414, where the trained machine learning may be stored. ... If not, process 2500 may return to operation 2508, where the model is trained again, having its weights, biases, and other hyperparameters adjusted based on results of an optimization function").
Claims 22-23 are rejected under 35 U.S.C. 103 as being unpatentable over Zhou et al. (US 20230329646 A1), hereinafter Zhou, in view of Pahlevan et al. (US 20230420132 A1), hereinafter Pahlevan, and further in view of Aykut et al. (US 20230196567 A1), hereinafter Aykut, and Ahmed, M Waqar (“Understanding Mean Absolute Error (MAE) in Regression: A Practical Guide.” Medium, 24 Aug. 2023, medium.com/@m.waqar.ahmed/understanding-mean-absolute-error-mae-in-regression-a-practical-guide-26e80ebb97df).
Regarding claim 22, the system of claim 20 is obvious over Zhou and Pahlevan, as explained above. Neither Zhou nor Pahlevan explicitly discloses that the metric comprises a mean absolute error between the physiological state prediction and the target physiological state.
However, Aykut teaches devices, systems, and methods for providing real-time, non-invasive monitoring of one or more physiological parameters of a patient (Abstract) wherein the metric comprises a mean absolute error between the physiological state prediction and the target physiological state (paragraph [0111], "the predicted heart rate of an illustrative SSQ-Unet model has a mean absolute error (MAE) of 0.95 beats per minute (BPM)").
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 Zhou and Pahlevan with the teachings of Aykut so that the metric comprises a mean absolute error between the physiological state prediction and the target physiological state, because the mean absolute error is less sensitive to outliers, easy to interpret, and straightforward to calculate and understand, as taught by Ahmed.
Regarding claim 23, the system of claim 22 is obvious over Zhou, Pahlevan, Aykut, and Ahmed, as explained above. Zhou further discloses that the target physiological state is based on annotated training data indicating expected physiological state values (paragraph [0128], "Generating synthetic acoustic data with a preserved class label greatly expands the training data available for acoustic classification models to learn from"; paragraph [0218], "The recorded sounds are classified as either patent or stenosed based on duplex ultrasound findings, which validates the data and provides the ground truth label. The sounds are converted into respective image representations in the manner previously described. Additional image representations can be generated using one or more of the data augmentation techniques disclosed herein, such as masking, GAN, horizontal flipping, etc. The image representations (both real and synthetic) are used to train a CNN for classification"; paragraph [0303], "The disclosed models are trained using supervised learning with labels validated from concurrent duplex ultrasound").
Claim 24 is rejected under 35 U.S.C. 103 as being unpatentable over Zhou et al. (US 20230329646 A1), hereinafter Zhou, in view of Odame et al. (US 12702323 B2), hereinafter Odame.
Regarding claim 24, Zhou discloses the system of claim 19, as explained above. Zhou further discloses that a first branch receives a pairing of Mel spectrogram and power spectral density features (paragraph [0113], "Referring initially to FIG. 8A, a biomedical acoustics classifier 800 utilizing a spectrogram and a power spectrum is shown").
Zhou does not explicitly disclose that a second branch receives a pairing of Mel frequency cepstral coefficient and root mean square features.
However, Odame teaches method for autonomous detection of asthma symptoms and inhaler use (column 2, lines 43-44), wherein a neural network receives a pairing of Mel frequency cepstral coefficient and root mean square features (column 5, lines 35-39, "firmware 160 and model 162 first extract special features relating to the temporal (e.g., RMS energy), spectral (e.g., MFCCs)...characteristics of the relevant signals 124. Firmware 160 then utilizes one or more signal processing (such as template matching), statistical inferencing (e.g., Bayesian methods) and/or pattern recognition/machine learning techniques (e.g., SVM, HMM, DNN classifiers), to qualify and generate events 182").
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 Zhou with the teachings of Odame so that a second branch receives a pairing of Mel frequency cepstral coefficient and root mean square features, because doing so produces extracted features that can be used to generate events with a probability, likelihood, and/or confidence that the event is a particular symptom (Odame, column 5, lines 46-49).
Claim 27 is rejected under 35 U.S.C. 103 as being unpatentable over Zhou et al. (US 20230329646 A1), hereinafter Zhou, in view of Iyer et al. (US 20210321890 A1), hereinafter Iyer.
Regarding claim 27, Zhou discloses the system of claim 25, as explained above. Zhou does not explicitly disclose that the neural network has been trained to generate output nodes corresponding to different heart-rate values within a specified range.
However, Iyer teaches a system for detecting fetal movement by analyzing a fetal heart rate to identify a fetal heart rate acceleration (paragraph [0015]), wherein the neural network has been trained to generate output nodes corresponding to different heart-rate values within a specified range (paragraph [0023], "an output layer of the ANN model is a softmax layer that has a neuron node for each heart rate value").
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 Zhou with the teachings of Jones so that the neural network has been trained to generate output nodes corresponding to different heart-rate values within a specified range, because doing so enables the system to only analyze values that meet a certain quality threshold, and are therefore more accurate (Iyer, paragraphs [0022]-[0023]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Tsai et al. (US 20160354053 A1) discloses a system for recognizing physiological sound using machine learning techniques
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
/REX R HOLMES/Primary Examiner, Art Unit 3796
Prosecution Insights