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 .
Priority
Current application, US Application No. 17/979,860 filed on 11/03/2022, claims foreign priority to TW 111115700 filed on 04/25/2022.
Examiner acknowledges that the certified copy of foreign priority document has been received. However, the certified English translation copy of the original foreign document, which is not written in English, has not been received. There is no requirement to submit certified English translation copy at this stage according to 37 CFR 1.55(g)(3). However, should the need of certified English translated copy arise according to the cases mentioned in 37 CFR 1.55(g)(3), submission may be requested in the future.
DETAILED ACTION
This office action is responsive to the amendment filed on 06/25/2026. Claims 1-2, 5-7, 9-11 and 14-17 are currently pending. Claims 3-4, 8 and 12-13 are canceled by applicant’s request.
Response to Amendment
Applicant's amendment is entered into further examination and appreciated by the examiner.
Response to Arguments/Remarks
Regarding remarks on the objections to the claims, the amendment is accepted and the previous objections are withdrawn.
Regarding remarks on the rejections under 35 USC 112(b), the amendment could not cure all the ambiguities so newly updated rejections are issued. See the updated office action below.
Regarding remarks on the rejections under 35 USC 103, a high level of ambiguities prevents a new reference from being applied. Until the rejections under 35 USC 112(b) being overcome, existing rejections are maintained.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1-2, 5-7, 9-11 and 14-17 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention. As per claim 1, claim recites limitations as below:
a computing device, the computing device comprising an abnormality detection model trained with a large number of cycle variable frequency signals (CVF signals for short) which underwent time-frequency transform and the anomaly detection model uses a denoising convolutional autoencoder model for signal reconstruction, (1.A)
wherein the computing device is signal-connected to the signal sensor to receive the sample signal to be tested, perform a correction on the sample signal to be tested, and perform a time-frequency transform on a one-dimensional signal generated after the correction to generate a two-dimensional time-frequency signal, the computing device reconstructs the two-dimensional time-frequency signal by using an abnormality detection model to calculate a reconstructed difference value which is a difference value between the two-dimensional time-frequency signal being inputted into the abnormality detection model and an output of the abnormality detection model, the output comprising a reconstruction of the two-dimensional time-frequency signal inputted into the abnormality detection model by applying change features of each frequency and time learned by the model through training to the reconstruction, (1.B)
The limitation (1.A) is ambiguous because the first model called “the anomaly detection model” treats the second model “a denoising convolutional autoencoder model” as a different model than the first one. However, the specification discloses the two models are a single identical model (see specification - the abnormality detection model 16 in the computing device 14 is a denoising convolutional autoencoder model [0024]).
Besides, the limitation (1.B) is ambiguous because the limitation “an output of the abnormality detection model, the output comprising a reconstruction of the two-dimensional time-frequency signal inputted into the abnormality detection model” is not consistent with the limitation (1.A) regarding the output of the abnormality detection model. Furthermore, the phrase “a reconstruction of two-dimensional time-frequency signal” in the limitation “a reconstruction of the two-dimensional time-frequency signal inputted into the abnormality detection model” is illogical because the “two-dimensional time-frequency signal” did not go through any transformation which would require reconstruction although the signal may have gone through a modification or a calibration.
As can be seen below, the limitations (1.A) and (1.B) are not consistent in interpreting the output of the abnormality detection model.
PNG
media_image1.png
493
1027
media_image1.png
Greyscale
PNG
media_image2.png
478
1017
media_image2.png
Greyscale
For the sake of examination, the limitations (1.B) and (1.C) should be amended reflecting the following graphical interpretation of limitation (1.B’) by Examiner.
PNG
media_image3.png
499
995
media_image3.png
Greyscale
As per claim 10, the limitations “reconstructing the two-dimensional time-frequency signal by using an abnormality detection model to calculate a reconstructed difference value” and “the anomaly detection model uses a denoising convolutional autoencoder model for signal reconstruction, wherein the reconstructed difference value comprises a difference value between the two-dimensional time-frequency signal being inputted into the abnormality detection model and an output of the abnormality detection model, the output comprising a reconstruction of the two-dimensional time-frequency signal inputted into the abnormality detection model” are ambiguous as explained in claim 1 above.
As per claims 2, 5-7, 9, 11 and 14-17, claims are also rejected because base claims 1 and 10 are rejected.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 2, 5-6, 9, 10-11, 14-15 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Xiong (CN 111800811 B) in view of Bin (US 20220279300 A1), Dowling (US 5799114 A) and Kayser (Kayser, Mike, and Victor Zhong. "Denoising convolutional autoencoders for noisy speech recognition." CS231 Standford Reports (2015)) best understood by the examiner.
As per claim 1, Xiong discloses
A signal abnormality detection system (detection method, abnormality, target signal [abs], intelligent … detection system [pg. 2 line 38], abnormality detection method, system automatically detect the occurrence of abnormal condition [pg. 3 line 4-9]) , comprising:
a signal sensor, generating a sample signal to be tested through sensing; (a signal obtaining module for obtaining the target signal to be detected [pg. 5 line 21], test data set, testing, test sample [pg. 9 line 2-20])
However, Xiong is silent regarding the signal sensor comprising a microphone or an accelerometer, and the sample signal to be tested comprising a cycle variable frequency signal.
Bin discloses the signal sensor comprising a microphone or an accelerometer, and the sample signal to be tested comprising a cycle variable frequency signal (audio signal [0010], side note: audio signal is a cycle variable frequency signal, classifier, model [0072], microphone, sensors … accelerometer [0181]).
Bin is in the same art concerning the signal quality using a model generating two-dimensional time-frequency signal as Xiong (see Bin - audio quality [0010], spectrogram [0054], machine learning classifier, training data, extracted features [0067-0070], classifier, model [0072]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time when invention is filed before the effective filing date of the current application to modify the teachings of Xiong in view of Bin to use the signal sensor comprising a microphone or an accelerometer, and the sample signal to be tested comprising a cycle variable frequency signal in order to enhance an analysis accuracy of the obtained signal.
Xiong further discloses
and a computing device, the computing device comprising a signal reconstruction model (mode, neural network, autoencoder [pg. 8 line 1-35])
wherein the computing device is signal-connected to the signal sensor to receive the sample signal to be tested, (electronic device, computer program, processor [pg. 6 line 1-9], obtaining the target signal to be detected [pg. 6 line 12])
However, Xiong is silent regarding performing a correction on the sample signal. (see specification for a scaling operation – [0026, Figs. 3A-3C]).
Dowling discloses adjusting and aligning the sample signal with the predefined mathematical operator by dilating and translating (stably identifying and characterizing a transient contained in a sampled set of data, analyzing the first data set using a predefined mathematical operator to produce a second data set, adjusting the location of the discrete data pattern in the second data set to align it with the predefined mathematical operator in accordance with the measured location difference [abs, col 3 line 42-67], Wavelets, dilations, translations, wavelets, time scale plane, inner product [col 2 line 32-52, eq. 1 and 2], wavelet decomposition [col 4 line 9 – 14, Fig, 2a-2h], the analyzing function is the discrete wavelet transform. and the alignment of the sample data points with the transient is arbitrary [col 10 line 31- col 12 line 57, Fig. 4-6]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time when invention is filed before the effective filing date of the current application to modify the teachings of the combined prior art in view of Dowling to perform a correction on the sample signal to be tested, and perform a time-frequency transform on a one-dimensional signal generated after the correction to generate a two-dimensional time-frequency signal in order to enhance an analysis accuracy of the obtained signal.
Xiong continues to disclose
perform a time-frequency transform on a one-dimensional signal generated after the correction to generate a two-dimensional time-frequency signal, (converting the signal into the corresponding time-frequency image [abs, pg. 2 line 5-6], generating the corresponding original time-frequency image according to the time frequency image of the target signal [pg. 3 line 13-14], short-time Fourier transform ‘STFT’ [pg. 9 line 5-20, pg. 14 line 16-20, claim 4])
the computing device reconstructs the two-dimensional time-frequency signal by using an abnormality detection model to calculate a reconstructed difference value, (inputting the original time-frequency image into the pre-created generated countermeasure network, equivalent to the abnormality detection model, obtaining the reconstructed time-frequency image corresponding to the original time-frequency image; using the original time-frequency image and the reconstructed time-frequency image to determine the abnormal value [pg. 3 line 13-29])
However, the combined prior art is silent regarding a reconstructed difference value
which is a difference value between the two- dimensional time-frequency signal being inputted into the signal reconstruction model and an output of the signal reconstruction model, the output comprising a reconstruction of the two-dimensional time-frequency signal inputted into the signal reconstruction model by applying change features learned by the model through training to the reconstruction.
Kayser discloses using a trained reconstruction model to compute the reconstructed difference value which is a difference value between the two- dimensional time-frequency signal being inputted into the signal reconstruction model and an output of the signal reconstruction model (spectrogram, convolutional autoencoder, train the model [pg. 4 right col par 3-1 from the bottom], convolutional autoencoder, reconstruction error [abs], reconstruction, spectrogram of clean audio [pg. 2 right col par 1- pg. 3 left col par 2], reconstruction error [pg. 4 left col par 2 – pg. 5 left col par 3]).
Kayser is in the same art of analyzing the signal quality using a model as the combined prior art.
Therefore, it would have been obvious to one of ordinary skill in the art at the time when invention is filed before the effective filing date of the current application to modify the teachings of the combined prior art in view of Kayser to use the signal sensor comprising a microphone or an accelerometer, and the sample signal to be tested comprising a cycle variable frequency signal in order to enhance an analysis accuracy of the obtained signal.
Xing further discloses
and the computing device performs comparison to determine whether the reconstructed difference value is greater than a detection threshold to determine whether the sample signal to be tested is an abnormal sample. (generating the abnormal detection result of the target signal according to the abnormal value and the threshold value [pg. 3 line 19-20]).
As per claim 10, Xiong discloses
A signal abnormality detection method, (detection method, abnormality, target signal [abs], abnormality detection method, system automatically detect the occurrence of abnormal condition [pg. 3 line 4-9])
Xiong in view of Dowling discloses the remaining limitations as shown in claim 1 above.
As per claims 2 and 11, Xiong and Dowling disclose claims 1 and 10 set forth above.
Dowling further discloses
calculating an inner product of each interval of the sample signal to be tested according to a standard sample signal and cutting and retaining a signal in an interval corresponding to the largest inner product as the one-dimensional signal. (inner product, dilations, translations, wavelets, time scale plane, [col 2 line 32-52, eq. 1 and 2], wavelet decomposition [col 4 line 9 – 14, Fig, 2a-2h], the analyzing function is the discrete wavelet transform and the alignment of the sample data points with the transient is arbitrary [col 10 line 31- col 12 line 57, Fig. 4-6])
As per claims 5 and 14, Xiong and Dowling disclose claims 1 and 10 set forth above.
Xiong discloses
a training method of the abnormality detection model comprises: adding random noise to a part of a plurality of corrected one-dimensional normal sample signals, and performing the time-frequency transform on the one-dimensional normal sample signals to separately generate a plurality of pieces of two-dimensional time-frequency training data and a plurality of pieces of two-dimensional time-frequency test data; (determining the training data set and the test data set, noise interference … random, STFT [pg. 9 line 2-20])
performing model training on an initial model by using the plurality of pieces of two-dimensional time-frequency training data, and optimizing a model parameter to construct the abnormality detection model; (building neural network, initial generation … the network, [pg. 9 line 21 – pg. 11 line 11], training neural network, update the weight of the encoder [pg. 11 line 12 – pg. 12 line 29],
and inputting the plurality of pieces of two-dimensional time-frequency test data into the abnormality detection model to calculate a difference value between an input and an output and setting the largest difference value as the detection threshold. (calculating the threshold value, inputting each original test time-frequency image in the training data into the generated countermeasure network, … test abnormal value, [pg. 13 line 1 – 14], threshold value generating module [pg. 18 line 28-p pg. 19 line 19])
(judging whether the abnormal value is greater than the threshold value if so, the abnormal detection result of the target signal is not abnormal; if not, then the abnormal detection result of the target signal is abnormal [pg. 5 line 10-12], implying the largest difference should be used for determining a normal threshold)
As per claims 6 and 15, Xiong and Dowling disclose claims 5 and 14 set forth above.
Xiong further discloses wherein the initial model is a denoising convolutional autoencoder model. (autoencoder, AE [pg. 8 line 23-24], training data set, noise interference [pg. 9 line 2-20], building neural network, convolution layer [pg. 9 line 21 – 36]).
As per claims 9 and 17, Xiong and Dowling disclose claims 1 and 10 set forth above.
Xiong further discloses the time-frequency transform is a short time Fourier transform. (STFT [pg. 9 line 5-20, pg. 14 line 17]).
Claims 7 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Xiong and Dowling in view of Zhang (CN 105067129 A).
As per claims 7 and 16, Xiong and Dowling disclose claims 5 and 14 set forth above.
Dowling implies the signal to be a cycle variable frequency signal (isolated data patterns, or transients, in a signal [col 1 line 20], accelerometer, nature of impulse, detects and analyzes … natural frequency transients [col 13 29-39]), but fails to explicitly recite a signal to be tested is a cycle variable frequency signal.
Zhang recites a variable cycle frequency signal (variable-cycle frequency domain signal identification [abs]).
Dowling discloses calculating an inner product of each interval of the sample signal to be tested according to a standard sample signal and cutting and retaining a signal in an interval corresponding to the largest inner product as the one-dimensional signal. (inner product, dilations, translations, wavelets, time scale plane, [col 2 line 32-52, eq. 1 and 2], wavelet decomposition [col 4 line 9 – 14, Fig, 2a-2h], the analyzing function is the discrete wavelet transform and the alignment of the sample data points with the transient is arbitrary [col 10 line 31- col 12 line 57, Fig. 4-6])
Therefore, it would have been obvious to one of ordinary skill in the art at the time when invention is filed before the effective filing date of the current application to modify the teachings of the combined prior art in view of Zhang to include the sample signal to be tested being a cycle variable frequency signal, collect a plurality of normal cycle variable frequency signals, select a cycle variable frequency signal with a complete cycle from the plurality of cycle variable frequency signals as a standard sample signal, and respectively calculate an inner product of each interval of each of the remaining normal cycle variable frequency signals according to the standard sample signal, and cut and retain signals in intervals corresponding to the largest inner product as the one-dimensional normal sample signals for being able to accurately handle a variety of different signal types.
Notes with regard to Prior Art
The prior arts made of record and not relied upon are considered pertinent to applicant's disclosure.
Zhang (CN 105067129 A) discloses obtaining variable-cycle frequency domain signal ([abs[).
Chen (CN 114098764 B) recites use of the largest inner product (residual error, so that the base function with the largest absolute value of the inner product [pg. 24 line 23-26]).
Bouska (US 20180203142 A1) also recites use of the largest inner product (the wavelet, seismic sensor [0092], an inner product of the measured waveform at a selected time and sample waveforms having known periodic lengths can be calculated. The sample which provides the largest inner product has a periodic length which most closely matches the periodic length of the waveform measured by the sensor, and so that periodic length is said to be the measured periodic length, which is used to determine the scaling factor, a. [0162]).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for replying 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.
Contact Information
Any inquiry concerning this communication or earlier communications from the examiner should be directed to, whose telephone number is (408) 918-7569. The examiner can normally be reached on M, Th & F 8-5, T 2-7, and W 8-1.
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, Arleen M Vazquez can be reached on 571-272-2619. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DOUGLAS KAY/
Primary Examiner, Art Unit 2857
Prosecution Insights