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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 3/27/25 and 5/18/26 are being considered by the examiner.
Claim Objections
Claim 1 is objected to because of the following informalities: Applicant is encouraged to change “comprising” in line 2 to recite –comprising:-- so it is clear what elements are a part of the apparatus. Appropriate correction is required.
Claim 8 is objected to because of the following informalities: for the limitation “the trained first and second machine-learning-based classifier” in line 2, Applicant is encouraged to change the limitation to recite --the trained first and second machine-learning-based classifiers—since it is plural. Appropriate correction is required.
Claim 11 is objected to because of the following informalities: for the limitation “wherein the trained first, second, and third second machine-learning-based classifiers” in lines 1-2, Applicant is encouraged to recite --wherein the trained first, second, and third machine-learning-based classifiers—to remove the extra “second” recited. Appropriate correction is required.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that use the word “means” or “step” but are nonetheless not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph because the claim limitation(s) recite(s) sufficient structure, materials, or acts to entirely perform the recited function. Such claim limitation(s) is/are:
“training the first machine-learning-based classifier (16) by means of ground truth audio signals” in claim 15; and
“training the second machine-learning-based classifier (18) by means of ground truth audio signals” in claim 15.
Because this/these claim limitation(s) is/are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are not being interpreted to cover only the corresponding structure, material, or acts described in the specification as performing the claimed function, and equivalents thereof.
If applicant intends to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to remove the structure, materials, or acts that performs the claimed function; or (2) present a sufficient showing that the claim limitation(s) does/do not recite sufficient structure, materials, or acts to perform the claimed function.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception, specifically an abstract idea.
Step 1
The claimed invention in claims 1 and 14 are directed to statutory subject matter as the claims recite an apparatus and a method for classifying at least one audio signal.
Step 2A, Prong One
Regarding claims 1 and 14, the recited steps are directed to a mental process of performing concepts in a human mind or by a human using a pen and paper (see MPEP 2106.04(a)(2) subsection (III)).
Regarding claims 1 and 14, the limitations of “map the input information (22) of the audio signal to one of a first and a second class of audio signals (24; 26); if the audio signal (20) belongs to the first class of audio signals (24), map the input information (22) of the audio signal belonging to the first class of audio signals (24) to one of a plurality of third classes (28) of audio signals; and output information on which classes the audio signal (20) belongs to” are a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, these limitations are nothing more than a medical professional receiving a print out of an audio signal, mapping information of the audio signal into one of a first or a second class, and if the audio signal belongs to the first class, mapping information of the audio signal in the first class to one of a plurality of third classes, and writing down information on which classes the audio signal belongs to.
Step 2A, Prong Two
For claims 1 and 14, the judicial exception is not integrated into a practical application. In particular, claims 1 and 14 recite “an input interface (12) configured to receive input information (22) of the audio signal (20) and an output interface.” The input interface and output interface are recited at a high-level of generality and amount to nothing more than parts of a generic computer. Additionally, Applicant includes a trained first machine-learning-based classifier (16) and a trained second machine-learning-based classifier (18) which are nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application.
Step 2B
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into
a practical application, the additional elements of the input interface and output interface are recited at a high-level of generality and amount to nothing more than parts of a generic computer. Additionally, Applicant includes a trained first machine-learning-based classifier (16) and a trained second machine-learning-based classifier (18) which are nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Further, simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer functions that are well-understood, routine and conventional activities previously known to the industry, as discussed in Alice Corp., 573 U.S. at 225, 110 USPQ2d at 1984 (see MPEP § 2106.05(d)).
Regarding dependent claims 2-13 and 15, the limitations of claims 1 and 14 further define the limitations already indicated as being directed to the abstract idea.
Claims 2-4 further define the abstract idea in claim 1.
Regarding claim 5, the limitation of “extract, from the audio signal (20), a plurality of features (22) characterizing the audio signal as the input information (22) of the audio signal (20)” is a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, this limitation is nothing more than a medical professional extracting a plurality of features from the print out of the audio signal. The preprocessor (14) is recited at a high-level of generality and amounts to nothing more than a part of a generic computer.
Claim 6 further defines the abstract idea in claim 5. The preprocessor (14) is recited at a high-level of generality and amounts to nothing more than a part of a generic computer.
Regarding claims 7-8 and 11, Applicant includes further details about the machine learning, which is nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application.
Regarding claim 9, Applicant includes a trained third machine-learning-based classifier (42), which is nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application. The limitations of “determine if the audio signal (20) belongs to the first or to the second class of audio signals, wherein it is known beforehand which of the first ground truth audio signals are related to which of the first and second classes of audio signals; and determine if the audio signal (20) belongs to one of the plurality of third classes (28), wherein it is known beforehand which of the second ground truth audio signals are related to which of the third classes of audio signals” are a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, these limitations are nothing more than a medical professional using the print out of the audio signal to determine if it belong to the first or to the second class of audio signals, wherein it is known beforehand which of the first ground truth audio signals are related to which of the first and second classes of audio signals; and to determine if it belongs to one of the plurality of third classes (28), wherein it is known beforehand which of the second ground truth audio signals are related to which of the third classes of audio signals.
Regarding claim 10, Applicant includes a trained third machine-learning-based classifier (42), which is nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application. The limitation of “map the input information (22) of an audio signal belonging to any one of the plurality of third classes (28) of audio signals and fulfilling an additional criterion to one of a plurality of fourth classes of audio signals” is a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, this limitation is nothing more than a medical professional using the print out of the audio signal to map it to any one of the plurality of third classes and fulfilling an additional criterion to one of a plurality of fourth classes of audio signals.
Regarding claim 12, Applicant includes further details about the machine learning, which is nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application. The limitation of “determine if the audio signal (20) belongs to one of the plurality of fourth classes (28), wherein it is known beforehand which of the third ground truth audio signals are related to which of the fourth classes of audio signals” is a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, this limitation is nothing more than a medical professional using a print out of the audio signal to determine if it belongs to one of the plurality of fourth classes (28), wherein it is known beforehand which of the third ground truth audio signals are related to which of the fourth classes of audio signals.
Claim 13 further defines the abstract idea in claim 10.
Regarding claim 15, Applicant includes further details about the machine learning, which is nothing more than the computer implementation/automation of an abstract mental process of screening a patient, which is what a physician typically does with a patient in a diagnostic setting. Merely including instructions to implement an abstract idea on a computer does not integrate a judicial exception into practical application. The limitations of “determine if the audio signal (20) belongs to the first or to the second class of audio signals; and determine if the audio signal belongs to one of the plurality of third classes (28), wherein it is known beforehand which of the ground truth audio signals are related to which of the third classes of audio signals” are a process, as drafted, covers performance of the limitation that can be performed by a human mind (including an observation, evaluation, judgment, opinion) under the broadest reasonable standard. For example, this limitation is nothing more than a medical professional using a print out of the audio signal to determine if it belong to the first or to the second class of audio signals, and to determine if it belongs to one of the plurality of third classes (28), wherein it is known beforehand which of the ground truth audio signals are related to which of the third classes of audio signals.
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 1-3, 5-9, and 14-15 are rejected under 35 U.S.C. 103 as being unpatentable over Yan (CN 111759345 published on 10/13/20 as cited in the IDS), and a copy relied upon is being furnished with this Office Action in view of Parvaneh (WO 2018046595 filed on 9/7/17 as cited in the IDS).
Regarding claims 1 and 14, Yan teaches an apparatus (10) and a method for classifying at least one audio signal (20), comprising receiving input information (22) of the audio signal (20) (page 9, ¶7-heart sounds collected in different auscultation areas to obtain the analysis results; page 5, ¶2-when step S201 is performed, the electronic stethoscope may be used to collect heart sounds in the aortic valve auscultation area, the mitral valve auscultation area, and the tricuspid valve auscultation area); a trained first machine-learning-based classifier (16) configured to map the input information (22) of the audio signal to one of a first and a second class of audio signals (24; 26) (page 8, ¶6-when step S302 is performed, the time spectrum of each heart sound is used as the acoustic feature, and the identification of the auscultation area of each heart sound is added, and they are jointly input into the convolutional neural network, so that the convolutional neural network can obtain the information of the auscultation area corresponding to each heart sound. Therefore, the first result that the heart sound is normal or abnormal is output on the output layer of the corresponding auscultation area; page 3, 2nd to last ¶-a multi-task learning convolutional neural network is used to classify the heart sounds collected in the aortic valve auscultation area, mitral valve auscultation area, and tricuspid valve auscultation area, and the output result is normal or abnormal heart valves); a trained second machine-learning-based classifier (18) configured to, if the audio signal (20) belongs to the first class of audio signals (24), map the input information (22) of the audio signal belonging to the first class of audio signals (24) to one of a plurality of third classes (28) of audio signals (page 3, 2nd to last ¶-for the heart sounds in the tricuspid auscultation area, since the corresponding tricuspid valve abnormality is only regurgitation, the second stage of analysis is not required. For the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area, if they are analyzed as abnormal by the convolutional neural network, the second stage of analysis is required. Therefore, in the second stage, the logistic regression hidden semi-Markov model can be used to segment the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area identified as abnormal in the first stage, and the segmented heart sounds can be divided into various stages. Energy parameters are used as features, and support vector machines are used for analysis. Through the analysis of support vector machine, the abnormal heart sounds collected in the aortic valve auscultation area can be analyzed for aortic valve stenosis or aortic regurgitation, and the abnormal heart sounds collected in the mitral valve auscultation area can be analyzed as mitral valve stenosis or secondary Cusp regurgitation); and outputting information on which classes the audio signal (20) belongs to (page 9, ¶2-when step S305 is performed, the above 6 energy features are input to the trained support vector machine, and the abnormal analysis result is output; page 3, 2nd to last ¶-a multi-task learning convolutional neural network is used to classify the heart sounds collected in the aortic valve auscultation area, mitral valve auscultation area, and tricuspid valve auscultation area, and the output result is normal or abnormal heart valves; page 8, ¶1-in the output results, the abnormal heart sounds collected in the aortic valve auscultation area are divided into two categories: aortic valve stenosis and/or aortic regurgitation, and the abnormal heart sounds collected in the mitral valve auscultation area are also divided into two categories: mitral Valve stenosis and/or mitral regurgitation).
However, Yan does not explicitly teach an input interface (12) and an output interface.
Parvaneh relates to systems, devices and methods for the detection of abnormal heart sounds (¶1). Parvaneh further teaches the invention using the following steps:
an input interface (12) (¶88-the user interface 44 may include a display, a mouse, and a keyboard for receiving user commands) and an output interface (¶70-one or more output devices 90 including, but not limited to, a monitor (e.g., of a workstation, a mobile device), a printer, a visual indicator (e.g., an LED assembly) and an audio indicator (e.g., a speaker)).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include an input interface (12) and an output interface of an input interface (12) and an output interface of Parvaneh in order to enable communication with a user such as an administrator (Parvaneh, ¶88), receive user commands (Parvaneh, ¶88), and report the final abnormality classification decision to a clinician (Parvaneh, ¶70).
Regarding claim 2, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 1, wherein the audio signal (20) comprises a plurality of cycles of a heart sound (Yan, page 5, ¶5-the heart sound signals of an integer number of cardiac cycles are marked as normal or abnormal, including aortic valve stenosis, aortic valve regurgitation, mitral valve stenosis, mitral valve regurgitation, and tricuspid valve regurgitation results; page 7, last ¶-an integer number of cardiac cycles is selected in each heart segment).
Regarding claim 3, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 2, wherein the first class (24) of audio signals denotes a pathological heart murmur (Yan, page 3, 2nd to last ¶-through the analysis of support vector machine, the abnormal heart sounds collected in the aortic valve auscultation area can be analyzed for aortic valve stenosis or aortic regurgitation, and the abnormal heart sounds collected in the mitral valve auscultation area can be analyzed as mitral valve stenosis or secondary Cusp regurgitation), and the second class (26) of audio signals denotes a healthy heart sound (Yan, page 3, 2nd to last ¶-in the first stage, a multi-task learning convolutional neural network is used to classify the heart sounds collected in the aortic valve auscultation area, mitral valve auscultation area, and tricuspid valve auscultation area, and the output result is normal or abnormal heart valves).
Regarding claim 5, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 1, comprising a preprocessor (14) configured to extract, from the audio signal (20), a plurality of features (22) characterizing the audio signal as the input information (22) of the audio signal (20) (Yan, page 8, ¶1-carry out low-pass and high-pass filtering on the heart sound; page 1, last ¶-short-time Fourier transform is performed on each heart sound to obtain the time spectrum of each heart sound; page 8, 2nd to last ¶-the Hilbert envelope, homomorphic envelope, wavelet envelope, and power spectral density spectrum are extracted from the heart sounds).
Regarding claim 6, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 5, wherein the preprocessor (14) is configured to extract time-domain features and/or frequency-domain features characterizing the audio signal (20) (Yan, page 8, ¶1-carry out low-pass and high-pass filtering on the heart sound; page 1, last ¶-short-time Fourier transform is performed on each heart sound to obtain the time spectrum of each heart sound; page 8, 2nd to last ¶-the Hilbert envelope, homomorphic envelope, wavelet envelope, and power spectral density spectrum are extracted from the heart sounds).
Regarding claim 7, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 1, wherein the trained first machine-learning-based classifier (16) is configured to implement a trained first boosting algorithm and/or wherein the trained second machine-learning-based classifier (18) is configured to implement a trained second boosting algorithm (Parvaneh, ¶103-feature -based classification S62 involves an implementation of a AdaBoost- abstain classifier. Specifically, AdaBoost is an effective machine learning technique for building a powerful classifier from an ensemble of .sup.Λ 'weak learners", whereby the boosted classifier ^i.sup.x) is modeled as a generalized additive model of many base hypotheses; ¶104; ¶114; ¶116-117).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include wherein the trained first machine-learning-based classifier (16) is configured to implement a trained first boosting algorithm and/or wherein the trained second machine-learning-based classifier (18) is configured to implement a trained second boosting algorithm of Parvaneh in order to classify normal/abnormal heart sounds (Parvaneh, ¶123) and to build a powerful classifier from an ensemble of weak learners (Parvaneh, ¶103).
Regarding claim 8, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 1, wherein the trained first and second machine-learning-based classifier (16; 18) are of the same type and differ by different respective training signals (Yan, page 3, ¶2-the heart valve abnormality analysis method based on the convolutional neural network of the present application is that the two-stage analysis of heart valve abnormalities can improve the accuracy and reduce the risk of misdiagnosis; page 2, ¶1-the normal or abnormal diagnosis result of the echocardiogram corresponding to each segment of heart sound is used as an output, and the convolutional neural network is trained to obtain a trained convolutional neural network; page 5, ¶1-train the support vector machine separately for the aortic valve auscultation area and the mitral valve auscultation area; page 9, ¶7-for the heart sounds collected in the three auscultation areas, the first stage analysis only performed two classifications, and the classification results were normal or abnormal. Then the heart sounds collected from the aortic valve auscultation area and mitral valve auscultation area classified as abnormal are segmented separately, and then the second stage analysis is performed to obtain aortic valve stenosis and/or aortic regurgitation, mitral valve stenosis and / Or the analysis result of mitral regurgitation).
Regarding claim 9, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 8, wherein the first machine-learning-based classifier (16) is trained based on first ground truth audio signals comprising the first class (24) and second class (26) of audio signals to enable the first machine-learning-based classifier (16) to determine if the audio signal (20) belongs to the first or to the second class of audio signals (Yan, page 3, 2nd to last ¶-in the first stage, a multi-task learning convolutional neural network is used to classify the heart sounds collected in the aortic valve auscultation area, mitral valve auscultation area, and tricuspid valve auscultation area, and the output result is normal or abnormal heart valves; page 5, ¶1-take the normal or abnormal results labeled in the echocardiographic diagnosis report corresponding to each heart sound as output, and train the convolutional neural network; page 2, ¶1-the step of training the convolutional neural network, including: segmenting the heart sounds of different auscultation areas in the training set; page 5, last ¶-the time spectrum of each heart sound is input into the convolutional neural network as the acoustic feature, and the normal or abnormal result of the echocardiographic diagnosis of the subject corresponding to each heart sound is used as the output, and the convolutional neural network is trained), wherein it is known beforehand which of the first ground truth audio signals are related to which of the first and second classes of audio signals (Yan, page 5, ¶1-take the normal or abnormal results labeled in the echocardiographic diagnosis report corresponding to each heart sound as output, and train the convolutional neural network), and wherein the second machine- learning-based classifier (18) is trained based on second ground truth audio signals comprising the first class (24) but not the second class of audio signals to enable the second machine- learning-based classifier (18) to determine if the audio signal (20) belongs to one of the plurality of third classes (28) (Yan, page 3, 2nd to last ¶-for the heart sounds in the tricuspid auscultation area, since the corresponding tricuspid valve abnormality is only regurgitation, the second stage of analysis is not required. For the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area, if they are analyzed as abnormal by the convolutional neural network, the second stage of analysis is required. Therefore, in the second stage, the logistic regression hidden semi-Markov model can be used to segment the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area identified as abnormal in the first stage, and the segmented heart sounds can be divided into various stages. Energy parameters are used as features, and support vector machines are used for analysis. Through the analysis of support vector machine, the abnormal heart sounds collected in the aortic valve auscultation area can be analyzed for aortic valve stenosis or aortic regurgitation, and the abnormal heart sounds collected in the mitral valve auscultation area can be analyzed as mitral valve stenosis or secondary Cusp regurgitation; page 5, ¶1-S204, taking the envelope spectrum feature and power spectrum feature of the center tone of the training set as input, and K states labeled according to the cardiac cycle as output. The labeling is done by professional physicians based on the ECG signal, and the logistic regression hidden semi-Markov model is trained , Wherein the envelope spectrum features include homomorphic envelope, Hilbert envelope, wavelet envelope, and the power spectrum feature includes power spectral density spectrum; S205, extract the energy features of each state in the cardiac cycle, and combine the energy features Input the support vector machine, and train the support vector machine separately for the aortic valve auscultation area and the mitral valve auscultation area; page 2, ¶2-the step of training a hidden semi-Markov model of logistic regression, including: dividing the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area into K according to the ECG signal Count the duration of each state in the K states of the central sound of the training set; page 2, ¶2-the probability distribution of each state duration is used as a parameter to train the logistic regression hidden semi-Markov model to obtain the trained logistic regression hidden Semi-Markov model), wherein it is known beforehand which of the second ground truth audio signals are related to which of the third classes of audio signals (Yan, page 5, ¶1-the report contains the diagnosis results, including normal, aortic stenosis, and aortic valve Regurgitation, mitral valve stenosis, mitral valve regurgitation, and tricuspid regurgitation are used to construct a training sample set; S202).
Regarding claim 15, the combination of Yan and Parvaneh teaches the method of claim 14, further comprising, during a training phase, training the first machine-learning-based classifier (16) by means of ground truth audio signals comprising the first class (24) and second class (26) of audio signals, to enable the first machine-learning-based classifier (16) to determine if the audio signal (20) belongs to the first or to the second class of audio signals (Yan, page 3, 2nd to last ¶-in the first stage, a multi-task learning convolutional neural network is used to classify the heart sounds collected in the aortic valve auscultation area, mitral valve auscultation area, and tricuspid valve auscultation area, and the output result is normal or abnormal heart valves; page 5, ¶1-take the normal or abnormal results labeled in the echocardiographic diagnosis report corresponding to each heart sound as output, and train the convolutional neural network; page 2, ¶1-the step of training the convolutional neural network, including: segmenting the heart sounds of different auscultation areas in the training set; page 5, last ¶-the time spectrum of each heart sound is input into the convolutional neural network as the acoustic feature, and the normal or abnormal result of the echocardiographic diagnosis of the subject corresponding to each heart sound is used as the output, and the convolutional neural network is trained); and training the second machine-learning-based classifier (18) by means of ground truth audio signals comprising the first class (24) but not the second class of audio signals, to enable the second machine-learning-based classifier (18) to determine if the audio signal belongs to one of the plurality of third classes (28) (Yan, page 3, 2nd to last ¶-for the heart sounds in the tricuspid auscultation area, since the corresponding tricuspid valve abnormality is only regurgitation, the second stage of analysis is not required. For the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area, if they are analyzed as abnormal by the convolutional neural network, the second stage of analysis is required. Therefore, in the second stage, the logistic regression hidden semi-Markov model can be used to segment the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area identified as abnormal in the first stage, and the segmented heart sounds can be divided into various stages. Energy parameters are used as features, and support vector machines are used for analysis. Through the analysis of support vector machine, the abnormal heart sounds collected in the aortic valve auscultation area can be analyzed for aortic valve stenosis or aortic regurgitation, and the abnormal heart sounds collected in the mitral valve auscultation area can be analyzed as mitral valve stenosis or secondary Cusp regurgitation; page 5, ¶1-S204, taking the envelope spectrum feature and power spectrum feature of the center tone of the training set as input, and K states labeled according to the cardiac cycle as output. The labeling is done by professional physicians based on the ECG signal, and the logistic regression hidden semi-Markov model is trained , Wherein the envelope spectrum features include homomorphic envelope, Hilbert envelope, wavelet envelope, and the power spectrum feature includes power spectral density spectrum; S205, extract the energy features of each state in the cardiac cycle, and combine the energy features Input the support vector machine, and train the support vector machine separately for the aortic valve auscultation area and the mitral valve auscultation area; page 2, ¶2-the step of training a hidden semi-Markov model of logistic regression, including: dividing the heart sounds collected in the aortic valve auscultation area and the mitral valve auscultation area into K according to the ECG signal Count the duration of each state in the K states of the central sound of the training set; page 2, ¶2-the probability distribution of each state duration is used as a parameter to train the logistic regression hidden semi-Markov model to obtain the trained logistic regression hidden Semi-Markov model), wherein it is known beforehand which of the ground truth audio signals are related to which of the third classes of audio signals (Yan, page 5, ¶1-the report contains the diagnosis results, including normal, aortic stenosis, and aortic valve Regurgitation, mitral valve stenosis, mitral valve regurgitation, and tricuspid regurgitation are used to construct a training sample set; S202).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Yan in view of Parvaneh as applied to claim 2 above, and further in view of Agarwal (US 20210169442 filed on 2/20/19).
Regarding claim 4, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 2. However, the combination of Yan and Parvaneh does not teach wherein the plurality of third classes (28) of audio signals relate to different pathological heart murmur levels.
Agarwal teaches wherein the plurality of third classes (28) of audio signals relate to different pathological heart murmur levels (¶33-the method/system may also determine a measure of an intensity of a heart sound comprising a murmur, for example according to the Levine grading scale. This may comprise determining, for each of a plurality of times during the murmur, a set coefficients characterising the time series acoustic heart signal data over a plurality of frequency bands. The coefficients may be MFCC (mel-frequency cepstral coefficients) coefficients. The set of coefficients for each time may then be mapped to a murmur intensity value, for example according to a learned regression; ¶5-the method/system may further involve classifying the time series acoustic heart signal data, for example using a neural network, into a plurality of heart sound categories to provide time series sound category data).
Agarwal relates to electronic stethoscope systems and methods (¶1).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include wherein the plurality of third classes (28) of audio signals relate to different pathological heart murmur levels of Agarwal in order to determine a measure of an intensity of a heart sound comprising a murmur, for example according to the Levine grading scale (Agarwal, ¶33). The murmur intensity values may then be combined, for example by averaging, to determine the measure of the intensity of the murmur (Agarwal, ¶33).
Claims 10 and 12-13 are rejected under 35 U.S.C. 103 as being unpatentable over Yan in view of Parvaneh as applied to claim 1 above, and further in view of Wilshaw (WO 2021205152 filed on 4/6/21 as cited in the IDS).
Regarding claim 10, the combination of Yan and Parvaneh teaches the apparatus (10) of claim 1. However, the combination of Yan and Parvaneh does not teach a trained third machine-learning-based classifier (42) configured to map the input information (22) of an audio signal belonging to any one of the plurality of third classes (28) of audio signals and fulfilling an additional criterion to one of a plurality of fourth classes of audio signals.
Wilshaw teaches a trained third machine-learning-based classifier (42) configured to map the input information (22) of an audio signal belonging to any one of the plurality of third classes (28) of audio signals (page 3, lines 10-14-a method of training a model to predict stage B2 DMVD in a dog, the method comprising: (i) processing characteristic data relating to a dog using the model to output an output value, the characteristic data comprising…murmur intensity; page 13, lines 16-25-the model may be derived using a machine learning process. In a machine learning process, characteristic data associated with a dog that has stage B2 DVMD or does not have stage B2 DVMD may be processed by an untrained model using a set of starting conditions. The starting conditions may be determined randomly. Alternatively, one or more of the starting conditions may be predetermined. The output of the untrained model may be compared with the condition of the dog associated with the characteristic data and the processing by the untrained model may be adjusted based on the comparison. This process may be repeated until the output of the model is associated with an accurate prediction of whether a dog associated with a particular set of characteristic data has stage B2 DVMD. This process may be referred to as a training process; page 10, lines 7-9-murmur intensity may be assessed using a simplified scale, where murmurs are classified as soft, moderate, loud or thrilling based on audibility upon cardiac auscultation) and fulfilling an additional criterion to one of a plurality of fourth classes of audio signals (page 2, lines 5-6-a model based on the parameters can be used to generate an output value associated with the probability of a dog having stage B2 DMVD; page 13, lines 16-19-in a machine learning process, characteristic data associated with a dog that has stage B2 DVMD or does not have stage B2 DVMD may be processed; page 29, lines 16-19-murmur intensity showed that the likelihood of being in stage B2 was greater when murmurs were more audible, with the comparison between loud and thrilling murmurs being the only pairwise combination that did not significantly differ (Table 4c)).
Wilshaw relates to method of diagnosing stage B2 degenerative mitral valve disease (DMVD) in a dog, a method of determining the probability of a dog having stage B2 DMVD, a method of training a model to predict stage B2 DMVD in a dog, and a related computer program and system (page 1, lines 5-7).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include a trained third machine-learning-based classifier (42) configured to map the input information (22) of an audio signal belonging to any one of the plurality of third classes (28) of audio signals and fulfilling an additional criterion to one of a plurality of fourth classes of audio signals of Wilshaw in order to train a model to predict stage B2 DMVD in a dog (Wilshaw, page 1, lines 6-7).
Regarding claim 12, the combination of Yan, Parvaneh, and Wilshaw teaches the apparatus (10) of claim 10 wherein the third machine-learning-based classifier (42) is trained based on third ground truth audio signals of the first class (24) but not the second class of audio signals (Wilshaw, page 3, lines 10-14-a method of training a model to predict stage B2 DMVD in a dog, the method comprising: (i) processing characteristic data relating to a dog using the model to output an output value, the characteristic data comprising…murmur intensity; page 13, lines 16-25-the model may be derived using a machine learning process. In a machine learning process, characteristic data associated with a dog that has stage B2 DVMD or does not have stage B2 DVMD may be processed by an untrained model using a set of starting conditions. The starting conditions may be determined randomly. Alternatively, one or more of the starting conditions may be predetermined. The output of the untrained model may be compared with the condition of the dog associated with the characteristic data and the processing by the untrained model may be adjusted based on the comparison. This process may be repeated until the output of the model is associated with an accurate prediction of whether a dog associated with a particular set of characteristic data has stage B2 DVMD. This process may be referred to as a training process) and fulfilling the additional criterion to enable the third machine-learning-based classifier (42) to determine if the audio signal (20) belongs to one of the plurality of fourth classes (28) (Wilshaw, page 2, lines 5-6-a model based on the parameters can be used to generate an output value associated with the probability of a dog having stage B2 DMVD; page 13, lines 16-19-in a machine learning process, characteristic data associated with a dog that has stage B2 DVMD or does not have stage B2 DVMD may be processed; page 29, lines 16-19-murmur intensity showed that the likelihood of being in stage B2 was greater when murmurs were more audible, with the comparison between loud and thrilling murmurs being the only pairwise combination that did not significantly differ (Table 4c); page 3, lines 11-15-processing characteristic data relating to a dog using the model to output an output value, the characteristic data comprising…age, breed; page 6, line 23-a dog of a certain breed and/or age may be suspected to have DMVD)), wherein it is known beforehand which of the third ground truth audio signals are related to which of the fourth classes of audio signals (Wilshaw, page 13, lines 16-25-the model may be derived using a machine learning process. In a machine learning process, characteristic data associated with a dog that has stage B2 DVMD or does not have stage B2 DVMD may be processed by an untrained model using a set of starting conditions. The starting conditions may be determined randomly. Alternatively, one or more of the starting conditions may be predetermined. The output of the untrained model may be compared with the condition of the dog associated with the characteristic data and the processing by the untrained model may be adjusted based on the comparison. This process may be repeated until the output of the model is associated with an accurate prediction of whether a dog associated with a particular set of characteristic data has stage B2 DVMD. This process may be referred to as a training process).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include wherein the third machine- learning-based classifier (42) is trained based on third ground truth audio signals of the first class (24) but not the second class of audio signals and fulfilling the additional criterion to enable the third machine-learning-based classifier (42) to determine if the audio signal (20) belongs to one of the plurality of fourth classes (28), wherein it is known beforehand which of the third ground truth audio signals are related to which of the fourth classes of audio signals of Wilshaw in order to train a model to predict stage B2 DMVD in a dog (Wilshaw, page 1, lines 6-7).
Regarding claim 13, the combination of Yan, Parvaneh, and Wilshaw teaches the apparatus (10) of claim 10, wherein the additional criterion is based on an age and/or a breed of a mammal the audio signal (20) belongs to (Wilshaw, page 3, lines 11-15-processing characteristic data relating to a dog using the model to output an output value, the characteristic data comprising…age, breed; page 6, line 23-a dog of a certain breed and/or age may be suspected to have DMVD).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include wherein the additional criterion is based on an age and/or a breed of a mammal the audio signal (20) belongs to of Wilshaw in order to train a model to predict stage B2 DMVD in a dog (Wilshaw, page 1, lines 6-7).
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Yan in view of Parvaneh, and further in view of Wilshaw as applied to claim 10 above, and further in view of Yu (US 20210224611 filed on 4/2/21).
Regarding claim 11, the combination of Yan, Parvaneh, and Wilshaw teaches the apparatus (10) of claim 10. However, the combination of Yan, Parvaneh, and Wilshaw does not teach wherein the trained first, second, and third second machine-learning-based classifiers (16; 18; 42) are of the same type and differ by different respective training signals.
Yu teaches wherein the trained first, second, and third second machine-learning-based classifiers (16; 18; 42) are of the same type and differ by different respective training signals (¶6-multiple trained machine-learning classifiers; ¶14-the first machine-learning classifier includes a first trained classifier that is trained using a first set of labeled data items, the second and third machine-learning classifiers include second and third trained classifiers that are trained on a second set of labeled data items, different from the first set of labeled data items; ¶7-the output classifications generated by the first machine-learning classifier are further classified by a second machine-learning classifier trained to reclassify misclassified elements in the classifications produced by the first machine-learning classifier, a first ensemble classifier combines the classifications generated by the first, second and third machine-learning classifiers to generate first ensemble classified data items; ¶27-training machine-learning classifiers that classify the unclassified elements and that reclassify the misclassified elements using a second set of labeled data items, different from the first set of labeled data items, based on classifications of the second set of labeled data items generated by the first trained classifier; Fig. 2A).
Yu relates to machine learning and, in particular, to an ensemble learning system that is trained on testing errors (¶2).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Yan to include wherein the trained first, second, and third second machine-learning-based classifiers (16; 18; 42) are of the same type and differ by different respective training signals of Yu in order to increase the precision and recall of the classification performed by the initial classifier (Yu, ¶59).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20210294845: an audio signal classifier including a feature extractor to extract metadata from an audio signal, the metadata defining a plurality of features of the audio signal, the feature extractor to generate a feature vector including selected features of the audio signal, the selected features including a duration of the audio signal, and each selected feature having a feature value. A machine learning model trained to classify the audio signal as one of a plurality of audio signal classes based on the feature vector. The machine learning model to provide a plurality of class values based on the feature values, each class value corresponding to one of the plurality of audio signal classes, the plurality of class values together indicating the class of the audio signal (Abstract).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to LAURA HODGE whose telephone number is (571) 272-7101. The examiner can normally be reached M-F: 8:00 am-5:00 pm.
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.
/LAURA HODGE/Examiner, Art Unit 3792
Prosecution Insights