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 .
Status of Claims
In the response dated March 11, 2026, Applicant amended claims 1, 7-8, 14-5, and 18. Claims 19 and 20 are added Claims 5, and 10-12 are canceled. Claims 1-4, 9 and 13-20 are pending.
Priority
Acknowledgment is made of applicant’s claim for priority. The certified copy has been filed in parent Application No. KR10-2022-0121967, filed on September 26, 2022.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on September 25, 2023 is being considered by the examiner.
Response to Arguments
In response to the argument put forward in the amendment, Examiner will address them in the order they were presented.
Regarding page 8, Applicant’s arguments have been considered and the 112(f) interpretation has been withdrawn.
Regarding pages 9-12, Applicant’s arguments have been considered but are moot in view of the amended claim language.
Eligible Subject Matter - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-4, 9 and 13-20 recite eligible subject matter under 35 U.S.C. 101 because the claimed invention is directed to statutory subject matter and practically applies recited judicial exceptions. MPEP 2106 provides guidelines for determining Subject Matter Eligibility with respect to potential ineligible subject matter. These steps below will elaborate on how this patent discloses a technical solution to a technical problem.
Step 1
The claims recite subject matter within a statutory category as a process, machine, and/or article of manufacture.
Step 2A Prong One
Claim 1 states:
A method of determining an analysis required section for signal segments with variable window sizes by an electrocardiogram data processing server including at least one processor, the method comprising:
receiving an electrocardiogram signal;
determining an analysis requirement of a first signal segment of the electrocardiogram signal having a first window size using a decision model;
determining a second window size of a second signal segment following the first signal segment depending on the analysis requirement of the first signal segment;
determining the analysis requirement of the second signal segment having a second window size using the decision model;
classifying each of the first and second signal segment as an analysis required section or a section not requiring analysis based on the analysis requirement of the first and second signal segment; and
storing, in a memory, data about signal segments belonging to the analysis required section and data about signal segments belonging to the section not requiring analysis.
The broadest reasonable interpretation of these steps includes mental processes and/or organizing human activity because each bolded component can practically be performed by the human mind or with pen and paper. Other than reciting generic computer terms like memory, processor, server, decision model, and electrocardiogram signal, the claims do not preclude the bold-font portions from practically being performed in the mind. For example, but for the “using a decision model” language, “determining an analysis requirement of a first signal segment of the electrocardiogram signal having a first window size” in the context of this claim encompasses a healthcare professional focusing in on pertinent regions of an ECG signal amidst a diagnosis. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” or “Organizing Human Activity” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Therefore:
A method of determining an analysis required section for signal segments with variable window sizes … the method comprising:
determining an analysis requirement of a first signal segment of the electrocardiogram signal having a first window size
determining a second window size of a second signal segment following the first signal segment
determining the analysis requirement of the second signal segment having a second window size
as drafted, could lay out the method in which an electrophysiologist determines the regions of an ECG print to interpret for their differential diagnosis. Therefore, under the broadest reasonable interpretation, these steps include multiple abstract ideas that will be identified as a single abstract idea moving forward.
Independent claims 15 and 18 cover similar steps of receiving an electrocardiogram signal, determining various window sizes to signal using a decision model, and classifying these signal sections as in need of additional analysis. These claims fall under the same category of an abstract idea and follows the same rationale as claim 1.
Dependent claims recite additional subject matter which further narrows or defines the abstract idea embodied in the claims (such as claim 5, reciting particular aspects of how “setting the second window size to a value larger than the first window size upon determining that the analysis requirement of the first signal segment is false” may be performed in the mind but for recitation of generic computer components).
Dependent claims add additional elements to their parent claims which will be further inspected in the following steps for a practical application to their abstract idea.
Step 2A Prong Two
This judicial exception of “Mental Processes” or “Organizing Human Activity” is integrated into a practical application. The independent claim’s additional elements integrate the abstract idea into a practical application because the additional elements improve the overall function of a computing system. To elaborate, the claims recite a combination of additional elements that improves the functioning of a computer, or an improvement to other technology or technical field. Specifically, the claims utilize hardware to automatically adjust window sizes of ECG signals and classify the quality of information to provide a specific improvement to the functioning of a computer or to any other technology/ technical field by minimizing the large computing times for noisy data, as described in MPEP 2106.05(a).
Dependent claims recite additional subject matter which further narrows and defines the abstract idea embodied in the claims by adding additional elements and practically applying the abstract idea. The dependent claims do not add additional abstract ideas and therefore do not prohibit subject matter eligibility.
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, 5-15, and 18-20 are rejected under 35 U.S.C. 103 as being unpatentable over Rubin et al. (US20200205687) in view of Nakagome et al. (US20210090740).
Regarding claim 1, Rubin teaches.
A method of determining an analysis required section for signal segments with variable window sizes by an electrocardiogram data processing server including at least one processor, the method comprising: ([Abstract] “The method includes: generating a time-frequency representation of an electrocardiogram (ECG) signal acquired over a time interval; processing the time-frequency representation” and [0026] “the electronic processor 20 may include a local processor of a workstation terminal and the processor of a server computer” and [0007]“At least one electronic processor is programmed” where the AF detection method comprises an analysis for segmenting window sizes to process electrocardiograms on servers)
receiving an electrocardiogram signal; ([0008] “acquiring an electrocardiograph (ECG) signal with an ECG measurement device;”)
determining an analysis requirement of a first signal segment of the electrocardiogram signal having a first window size using a decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index comprises the analysis requirement of a signal segment)
determining a second window size of a second signal segment following the first signal segment based on whether the analysis requirement of the first signal segment is true or false, ([0007] “generate a time-frequency representation of an ECG signal acquired over a time interval with values of a time dimension indexing time windows of a sliding time window over the ECG signal and with, for each indexed time window, values along a frequency dimension representing a frequency spectrum of the portion of the ECG signal in the indexed time window; process the time-frequency representation using a neural network (NN) ” where indexing time windows of a sliding time window comprises determining a second window size of a second segments following the first signal based on the timeframe [i.e., an analysis requirement])
determining the analysis requirement of the second signal segment having a second window size using the decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index comprises the analysis requirement of any signal segment; see optionally [007] “indexing time windows” where the analyzed signal segment comprises any time window)
classifying each of the first signal segment and the second signal segment as an analysis required section or a section not requiring analysis based on the analysis requirement of each of the first signal segment and the second signal segment; ([0031] “If the SQI is below the preselected SQI threshold, then the method 100 stops and the ECG signal is classified or determined to be noise. If the SQI is above the preselected SQI threshold, then the ECG signal is transmitted to the cloud computer processor 30 for further processing”)
Regarding claim 1, Rubin does not explicitly teach, as taught by Nakagome:
wherein, when the analysis requirement of the first signal segment is true, the second window size is determined to be identical to the first window size, wherein, when the analysis requirement of the first signal segment is false, the second window size is determined to be greater than twice the first window size, and wherein the second window size is determined based on the first window size of the first signal segment immediately preceding the second signal segment ([0050] “Extraction time windows having a plurality of mutually different periods are set from layer 0 to layer 3 in a period that is centered on the reference time t… windows 12 to 19 are set at mutually different positions on the time axis so as to be continuous without gaps” [0051] “The extraction time windows that are used differ depending on the feature values to be extracted. FFT processing is performed on the data that is extracted in accordance with the extraction time windows” where the processing of electrocardiogram signals alters which variably-sized time windows are utilized; see also [fig. 10], where the electrocardiogram time windows double in size as you change layer height yet are determined to be identically sized in the same layer, and where the time intervals may occur sequentially)
wherein, when the first signal segment is classified as the section not requiring analysis, ([0051] “The extraction time windows that are used differ depending on the feature values to be extracted.” Where determining feature vales to be extracted comprises classifying signal analysis requirements) the second signal segment is processed using an operation spanning a signal length at least twice a signal length of the first signal segment, ([fig. 10], where the electrocardiogram time windows double in size as you change layer height) wherein the determining of the analysis requirement for at least one of the first signal segment and the second signal segment is performed using a reduced number of operations relative to a fixed-window-based determination. ([0085] “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths”)
It would have been prima facie obvious to have modified Rubin with the teachings of Nakagome, with a reasonable expectation of success by explicitly creating timing windows that double based on the previous window size. Nakagome’s teaching would have taught Ruben that window setting can overcome common physiological signal computational limitations as “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths” [paragraph 0085]. Nakagome is adaptable to Rubin as both systems alter timing windows for electrocardiogram signal measurements in patients.
Regarding claim 6, Rubin-Nakagome as a combination teach all of the limitations of claim 1. Rubin also teaches:
further comprising setting, the second window size to a default value upon determining that the analysis requirement of the first signal segment is true. ([0034] “At 108, the cloud computer processor 30 is programmed to process the generated time-frequency representation using the NN 32 to output probabilities for rhythms of a set of rhythms. The set of rhythms can include one or more of… a noise rhythm or noisy recording (N). As used herein, the N rhythm constitutes an ECG signal that is too noisy to detect a particular rhythm, the rhythm is dominated by noise, and so forth.” And [0032] “the time window can be indexed by storing a start time (or a starting sample) of time window with a predefined length measured in time (or number of samples) and [0033] “Signal segments are then extracted from the spectrogram beginning at each of the detected QRS peaks. For each position of the sliding window (serving as the “x”-coordinate of the “image”), a spectrum is computed (serving as the “y”-coordinate of the “image”).” Where storing a predefined length measured in time comprises a default window size value and extracting detected QRS peaks while processing the time-frequency representation to output probabilities of rhythms comprises determining that the analysis requirement of a signal is true.)
Regarding claim 7, Rubin-Nakagome as a combination teach all of the limitations of claim 1. Rubin also teaches:
further comprising implementing the decision model to determine the first signal segment or the second signal segment as a section required analysis ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval; generating a time-frequency representation of an ECG signal acquired over a time interval with values of a time dimension indexing time windows of a sliding time window over the ECG signal and with, for each indexed time window, values along a frequency dimension representing a frequency spectrum of the portion of the ECG signal in the indexed time window” where indexing time windows comprise the first or second signal segment while implementing the signal quality index [comprising the decision model]) when a number of peaks, which is present within the first signal segment or the second signal segment and exceeds a peak reference value, exceeds a threshold. ([0021] “a signal quality assessment is performed to generate a signal quality index (SQI) indicative of noise in the ECG dataset. In an illustrative SQI formulation, QRS detection is performed to segment the dataset into individual heart beats which are compared with a template” where SQI formulation [i.e., decision model to determine each analysis requirement] references a QRS detection [i.e., based on a number of peaks]; see also [0033] “Signal segments are then extracted from the spectrogram beginning at each of the detected QRS peaks… the data set has some number of positions along the time dimension (note that successive time windows may overlap in time), with each position having some number of points in the frequency dimension storing the spectrum of the 0.25 second segment of ECG data.” Where the number of points comprises a peak reference value)
Regarding claim 8, Rubin-Nakagome as a combination teach all of the limitations of claim 7. Rubin also teaches:
further comprising further analyzing at least one signal segment classified as the analysis required section using an analysis model. ([0038] “At 212, if the SQI value is above the threshold, then a spectrogram of the ECG signal is generated. At 214, the spectrograms is analyzed by the NN 32, which in this embodiment is specifically a DenseNet, to assign probabilities of the ECG signal as AF, NSR, or O.” where the ECG signal [i.e., one signal segment] is classified using a neural network [comprising an analysis model] as Atrial Fibrillation, Normal Sinus Rhythm, etc. [i.e., analysis required section])
Regarding claim 9, Rubin-Nakagome as a combination teach all of the limitations of claim 8. Rubin also teaches:
wherein, the analysis model is learned by machine learning from labeled data. ([0039] “The NN 32 of the cloud computing processor 30 requires training. In general, the training employs a training set of ECG recordings which are labeled as to rhythm type (AF, NSR, O, or noise).”)
Regarding claim 13, Rubin-Nakagome as a combination teach all of the limitations of claim 8. Rubin also teaches:
wherein the electrocardiogram signal is a signal measured by a 1-channel measurement device. ([0007] “In another disclosed aspect, a device for detecting atrial fibrillation includes an ECG measurement device with one or more leads having one or more electrodes attachable to a patient.” Where one or more leads comprises a 1-channel measurement device)
Regarding claim 14, Rubin-Nakagome as a combination teach all of the limitations of claim 13. Rubin also teaches:
wherein the decision model is learned from data labeled with an attachment point of the electrocardiogram signal. ([0025] “the system 10 includes an electrocardiogram (ECG) measurement device 12 with at least one lead 14 defined between two electrodes 16 attachable to a patient” where defining a lead between two electrodes comprises data labeled with an attachment point of the electrocardiogram signal)
Regarding claim 15, Rubin teaches:
A server for determining an analysis required section for signal segments with variable window sizes, wherein the server comprises: ([0025] “an illustrative device or system 10 for performing an atrial fibrillation (AF) detection is shown.” and [0026] “the electronic processor 20 may include a local processor of a workstation terminal and the processor of a server computer”)
a communication interface configured to receive an electrocardiogram signal; ([0006] “a display device (24) to display the rhythm assigned to the ECG signal”)
a memory storing the electrocardiogram signal and data generated ([0026] “The workstation 18 can also include one or more databases or non-transitory storage media 26” and [0027] The illustrative workstation 18 is operatively connected with the ECG monitoring device 12 in order to receive an ECG data stream collected over a time period from the electrodes 16.”)
a processor configured to: ([0007] “At least one electronic processor is programmed”)
determine an analysis requirement of a first signal segment of the electrocardiogram signal having first window size using a decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index [i.e., a decision model] of a time interval [i.e., a signal having a first window size] comprises the analysis requirement of a signal segment)
determine a second window size of a second signal segment following the first signal segment based on whether the analysis requirement of the first signal segment is true or false; ([0007] “generate a time-frequency representation of an ECG signal acquired over a time interval with values of a time dimension indexing time windows of a sliding time window over the ECG signal ” where indexing time windows of a sliding time window comprises determining a second window size of a second segments following the first signal)
determine the analysis requirement of the second signal segment having second window size using the decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index comprises the analysis requirement of any signal segment; see optionally [007] “indexing time windows” where the analyzed signal segment comprises any time window)
classify each of the first signal segment and the second signal segment as the analysis required section or the section not requiring analysis based on the analysis requirement of the first and second signal segment; and ([0031] “If the SQI is below the preselected SQI threshold, then the method 100 stops and the ECG signal is classified or determined to be noise. If the SQI is above the preselected SQI threshold, then the ECG signal is transmitted to the cloud computer processor 30 for further processing”)
Regarding claim 1, Rubin does not explicitly teach, as taught by Nakagome:
wherein, when the analysis requirement of the first signal segment is true, the second window size is determined to be identical to the first window size, wherein, when the analysis requirement of the first signal segment is false, the second window size is determined to be greater than twice the first window size, and wherein the second window size is determined based on the first window size of the first signal segment immediately preceding the second signal segment ([0050] “Extraction time windows having a plurality of mutually different periods are set from layer 0 to layer 3 in a period that is centered on the reference time t… windows 12 to 19 are set at mutually different positions on the time axis so as to be continuous without gaps” [0051] “The extraction time windows that are used differ depending on the feature values to be extracted. FFT processing is performed on the data that is extracted in accordance with the extraction time windows” where the processing of electrocardiogram signals alters which variably-sized time windows are utilized; see also [fig. 10], where the electrocardiogram time windows double in size as you change layer height yet are determined to be identically sized in the same layer, and where the time intervals may occur sequentially)
wherein, when the first signal segment is classified as the section not requiring analysis, ([0051] “The extraction time windows that are used differ depending on the feature values to be extracted.” Where determining feature vales to be extracted comprises classifying signal analysis requirements) the second signal segment is processed using an operation spanning a signal length at least twice a signal length of the first signal segment, ([fig. 10], where the electrocardiogram time windows double in size as you change layer height) wherein the determining of the analysis requirement for at least one of the first signal segment and the second signal segment is performed using a reduced number of operations relative to a fixed-window-based determination. ([0085] “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths”)
It would have been prima facie obvious to have modified Rubin with the teachings of Nakagome, with a reasonable expectation of success by explicitly creating timing windows that double based on the previous window size. Nakagome’s teaching would have taught Ruben that window setting can overcome common physiological signal computational limitations as “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths” [paragraph 0085]. Nakagome is adaptable to Rubin as both systems alter timing windows for electrocardiogram signal measurements in patients.
Regarding claim 18, Rubin teaches:
A non-transitory, computer-readable storage medium storing instruction that, when executed by a processor, causes the processor to perform operations comprising: ([0006] “In one disclosed aspect, a non-transitory computer-readable medium stores instructions readable and executable by at least one electronic processor to perform an atrial fibrillation (AF) detection method.”)
receiving an electrocardiogram signal; ([0008] “acquiring an electrocardiograph (ECG) signal with an ECG measurement device;”)
determining an analysis requirement of a first signal segment of the electrocardiogram signal having first window size using a decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index [i.e., a decision model] of a time interval [i.e., a signal having a first window size] comprises the analysis requirement of a signal segment)
determining a second window size of a second signal segment following the first signal segment based on whether the analysis requirement of the first signal segment is true or false; ([0007] “generate a time-frequency representation of an ECG signal acquired over a time interval with values of a time dimension indexing time windows of a sliding time window over the ECG signal ” where indexing time windows of a sliding time window comprises determining a second window size of a second segments following the first signal)
determining the analysis requirement of the second signal segment having the second window size using the decision model; ([0008] “determining a signal quality index (SQI) of the ECG signal over the time interval;” where determining the signal quality index comprises the analysis requirement of any signal segment; see optionally [007] “indexing time windows” where the analyzed signal segment comprises any time window)
classifying each of the first signal segment and the second signal segment as an analysis required section or a section not requiring analysis based on the analysis requirement of the first signal segment and the second signal segment; and ([0031] “If the SQI is below the preselected SQI threshold, then the method 100 stops and the ECG signal is classified or determined to be noise. If the SQI is above the preselected SQI threshold, then the ECG signal is transmitted to the cloud computer processor 30 for further processing”)
Regarding claim 1, Rubin does not explicitly teach, as taught by Nakagome:
wherein, when the analysis requirement of the first signal segment is true, the second window size is determined to be identical to the first window size, wherein, when the analysis requirement of the first signal segment is false, the second window size is determined to be greater than twice the first window size, and wherein the second window size is determined based on the first window size of the first signal segment immediately preceding the second signal segment ([0050] “Extraction time windows having a plurality of mutually different periods are set from layer 0 to layer 3 in a period that is centered on the reference time t… windows 12 to 19 are set at mutually different positions on the time axis so as to be continuous without gaps” [0051] “The extraction time windows that are used differ depending on the feature values to be extracted. FFT processing is performed on the data that is extracted in accordance with the extraction time windows” where the processing of electrocardiogram signals alters which variably-sized time windows are utilized; see also [fig. 10], where the electrocardiogram time windows double in size as you change layer height yet are determined to be identically sized in the same layer, and where the time intervals may occur sequentially)
wherein, when the first signal segment is classified as the section not requiring analysis, ([0051] “The extraction time windows that are used differ depending on the feature values to be extracted.” Where determining feature vales to be extracted comprises classifying signal analysis requirements) the second signal segment is processed using an operation spanning a signal length at least twice a signal length of the first signal segment, ([fig. 10], where the electrocardiogram time windows double in size as you change layer height) wherein the determining of the analysis requirement for at least one of the first signal segment and the second signal segment is performed using a reduced number of operations relative to a fixed-window-based determination. ([0085] “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths”)
It would have been prima facie obvious to have modified Rubin with the teachings of Nakagome, with a reasonable expectation of success by explicitly creating timing windows that double based on the previous window size. Nakagome’s teaching would have taught Ruben that window setting can overcome common physiological signal computational limitations as “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths” [paragraph 0085]. Nakagome is adaptable to Rubin as both systems alter timing windows for electrocardiogram signal measurements in patients.
Regarding claim 19, Rubin-Nakagome as a combination teach all of the limitations of claim 1. Nakagome also teaches:
further comprising: determining a second threshold value to apply to the second signal segment ([0049] “The time window setter 120 sets the extraction time windows”) by multiplying a first threshold value by a ratio value of the second window size to the first window size, wherein the second threshold value is used to determine the analysis requirement for the second signal segment. ([Figure 10] and [0050] “FIG. 10 illustrates extraction time windows set with respect to a reference time t of an epoch In FIG. 10, time is represented on the horizontal axis. Extraction time windows having a plurality of mutually different periods are set from layer 0 to layer 3 in a period that is centered on the reference time t and that has a length of 256 sec… The time length of each of the windows 1 to 19 is set to a value corresponding to the re-sampling frequency of the equal interval RRIs” where figure 10 depicts the timing window length changing at a fixed ratio compared to the previous timing window with every change in layer)
It would have been prima facie obvious to have modified Rubin with the teachings of Nakagome, with a reasonable expectation of success by explicitly creating timing windows that double based on the previous window size. Nakagome’s teaching would have taught Ruben that window setting can overcome common physiological signal computational limitations as “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths” [paragraph 0085]. Nakagome is adaptable to Rubin as both systems alter timing windows for electrocardiogram signal measurements in patients.
Regarding claim 20, Rubin-Nakagome as a combination teach all of the limitations of claim 15. Nakagome also teaches:
further comprising: determining a second threshold value to apply to the second signal segment ([0049] “The time window setter 120 sets the extraction time windows”) by multiplying a first threshold value by a ratio value of the second window size to the first window size, wherein the second threshold value is used to determine the analysis requirement for the second signal segment. ([Figure 10] and [0050] “FIG. 10 illustrates extraction time windows set with respect to a reference time t of an epoch In FIG. 10, time is represented on the horizontal axis. Extraction time windows having a plurality of mutually different periods are set from layer 0 to layer 3 in a period that is centered on the reference time t and that has a length of 256 sec… The time length of each of the windows 1 to 19 is set to a value corresponding to the re-sampling frequency of the equal interval RRIs” where figure 10 depicts the timing window length changing at a fixed ratio compared to the previous timing window with every change in layer)
It would have been prima facie obvious to have modified Rubin with the teachings of Nakagome, with a reasonable expectation of success by explicitly creating timing windows that double based on the previous window size. Nakagome’s teaching would have taught Ruben that window setting can overcome common physiological signal computational limitations as “the calculation load is reduced by reducing the number of extraction time windows and setting longer time lengths” [paragraph 0085]. Nakagome is adaptable to Rubin as both systems alter timing windows for electrocardiogram signal measurements in patients.
Claims 2-4, and 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Rubin et al. (US20200205687) in view of Nakagome et al. (US20210090740) and further in view of Szabados et al. (US20210244339).
Regarding claim 2, Rubin-Nakagome as a combination teach all of the limitations of claim 1. Rubin does not explicitly teach, as taught by Szabados:
wherein, the decision model is a model learned with labeled data using a machine learning. ([0037] “computing system configured to infer a likelihood of an occurrence of cardiac arrhythmia by processing the output through a second subset of layers of the neural network, wherein the system is configured to train the neural network by: training a first neural network to identify a first feature by processing first training data of a first time period through the first neural network; freezing weights of the first neural network; training a second neural network to identify a second feature by processing second training data of a second time period through the first and second neural network, wherein the second time period is longer than the first time period; unfreezing weights of the first neural network; and training the first and second neural network simultaneously to identify the second feature by processing third training data of a third time period through the first and second neural network, wherein the third time period is longer than the first time period.”)
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 have modified Rubin with the teachings of Szabados, with a reasonable expectation of success, by incorporating labeled peak interval training data into neural network training to determine window size. This would have adapted the diagnosis patterns to each patient, creating more accurate results for disease states. Szabados is adaptable to Rubin as both inventions use machine learning to diagnose arrythmias and adjust window timing for electrocardiogram signals. Rubin would have found Szabados’s teaching while looking for methods to overcome industry’s data processing shortcomings as “if the computation is based only on the extracted sequence of R-peak locations (and not the whole ECG signal), estimating the AFib burden becomes a different algorithmic problem.” [0185].
Regarding claim 3, Rubin-Nakagome-Szabados as a combination teach all of the limitations of claim 2. Rubin also teaches
further comprising implementing the decision model to determine each analysis requirement of each signal segment based on a number of peaks that is present within each signal segment and that exceeds a peak reference value. ([0021] “a signal quality assessment is performed to generate a signal quality index (SQI) indicative of noise in the ECG dataset. In an illustrative SQI formulation, QRS detection is performed to segment the dataset into individual heart beats which are compared with a template” where SQI formulation [i.e., decision model to determine each analysis requirement] references a QRS detection [i.e., based on a number of peaks]; see also [0033] “Signal segments are then extracted from the spectrogram beginning at each of the detected QRS peaks… the data set has some number of positions along the time dimension (note that successive time windows may overlap in time), with each position having some number of points in the frequency dimension storing the spectrum of the 0.25 second segment of ECG data.” Where the number of points comprises a peak reference value)
Regarding claim 4, Rubin-Szabados as a combination teach all of the limitations of claim 3. Rubin does not explicitly teach, as taught by Szabados:
further comprising determining the peak reference value by learning the labeled data. ([0225] “In some embodiments, the neural network can be trained according to the current activity of the user. FIG. 22F is a schematic diagram of an embodiment for designing, training, and/or selecting a neural network based on… different types of peak sensitivities. There are certain activities that may require higher precision or accuracy to be able to effectively detect heart beat irregularities.”)
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 have modified Rubin with the teachings of Szabados, with a reasonable expectation of success, by incorporating labeled peak interval training data into neural network training to determine window size. This would have adapted the diagnosis patterns to each patient, creating more accurate results for disease states. Szabados is adaptable to Rubin as both inventions use machine learning to diagnose arrythmias and adjust window timing for electrocardiogram signals. Rubin would have found Szabados’s teaching while looking for methods to overcome industry’s data processing shortcomings as “if the computation is based only on the extracted sequence of R-peak locations (and not the whole ECG signal), estimating the AFib burden becomes a different algorithmic problem.” [0185].
Regarding claim 16, Rubin-Nakagome as a combination teach all of the limitations of claim 15. Rubin does not explicitly teach, as taught by Szabados:
wherein the decision model is a model learned with labeled data using machine learning. ([0225] “In some embodiments, the neural network can be trained according to the current activity of the user. FIG. 22F is a schematic diagram of an embodiment for designing, training, and/or selecting a neural network based on… different types of peak sensitivities. There are certain activities that may require higher precision or accuracy to be able to effectively detect heart beat irregularities.”)
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 have modified Rubin with the teachings of Szabados, with a reasonable expectation of success, by incorporating labeled peak interval training data into neural network training to determine window size. This would have adapted the diagnosis patterns to each patient, creating more accurate results for disease states. Szabados is adaptable to Rubin as both inventions use machine learning to diagnose arrythmias and adjust window timing for electrocardiogram signals. Rubin would have found Szabados’s teaching while looking for methods to overcome industry’s data processing shortcomings as “if the computation is based only on the extracted sequence of R-peak locations (and not the whole ECG signal), estimating the AFib burden becomes a different algorithmic problem.” [0185].
Regarding claim 17, Rubin-Szabados as a combination teach all of the limitations of claim 16. Rubin also teaches:
wherein the decision model is implemented to determine each analysis requirement of each signal segment based on a number of peaks that exists within each signal segment and that exceeds a peak reference value. ([0021] “a signal quality assessment is performed to generate a signal quality index (SQI) indicative of noise in the ECG dataset. In an illustrative SQI formulation, QRS detection is performed to segment the dataset into individual heart beats which are compared with a template” where SQI formulation [i.e., decision model to determine each analysis requirement] references a QRS detection [i.e., based on a number of peaks]; see also [0033] “Signal segments are then extracted from the spectrogram beginning at each of the detected QRS peaks… the data set has some number of positions along the time dimension (note that successive time windows may overlap in time), with each position having some number of points in the frequency dimension storing the spectrum of the 0.25 second segment of ECG data.” Where the number of points comprises a peak reference value)
Additional Considerations
The prior art made of record and not relied upon that is considered pertinent to applicant’s disclosure can be found on PTO-892 of the prior office action.
The following citations are cited as pertinent but not applied art in reference to this application.
Nakagome et al. (US20210090740) discloses a system in Figure 10 that doubles timing windows over a definite period of time.
Reiffman et al. (US20220151569) discloses a system for determining symptomatic conditions using ECG signal processing over varying timing windows.
Hamilton et al. (US20030208128) discusses displaying time windows.
Conclusion
THIS ACTION IS MADE FINAL. 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 ROBERT ANTHONY SKROBARCZYK whose telephone number is (571)272-3301. The examiner can normally be reached Monday thru Friday 7:30AM -5PM CST.
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, Unsu Jung can be reached at 571-272-8506. 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.
/R.A.S/Examiner, Art Unit 3792
/KAMBIZ ABDI/Supervisory Patent Examiner, Art Unit 3685
Prosecution Insights