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 .
DETAILED ACTION
Claims 1-2, 4-9, 11-16, and 18-21 are pending. Claims 1, 8, and 15 are independent.
This Application was published as US 20240331707.
Apparent priority is 3 April 2023.
The instant Application is directed to a method of detecting fraudulent voice authentication attempts and denying access if two voice samples meet a similarity threshold.
Response to Arguments
35 USC 103
Applicant’s arguments with respect to 35 USC 103 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
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.
Claim(s) 1-2, 4-5, 8-9, 11-12, 15-16, and 18-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Pilz (US 20100131279 A1) in view of SciPy ("find_peaks"), Marius (“Experimenting with acoustic fingerprinting for duplicate detection”), and James (“How to quantify spectrogram comparison”).
Regarding claim 1, Pilz discloses: 1. A system for digital voice data processing and authentication, the system comprising: at least one non-transitory storage device; and at least one processing device coupled to the at least one non-transitory storage device, (System Server 101 Fig. 4 – a server contains at least one storage device and one processing device.)
wherein the at least one processing device is configured to: receive a user interaction comprising a digital audio signal; capture a first audio segment and a second audio segment of the digital audio signal, the second audio segment subsequent the first audio segment wherein the first audio segment corresponds to a first instance of a word and the second audio segment corresponds to a second instance of the word, ("[0032] ... For this purpose he is instructed to input a Speech Sample 1, for instance to speak into the system a password assigned to him, and from this currently acquired speech sample, again in a training process, a current voice profile Speaker Model 1A is produced. Immediately thereafter the alleged user is instructed to input the same speech sample again, and this new sample is available as Speech Sample 1R. ..." )
wherein a natural language processing engine identifies the first audio segment and the second audio segment as two instances of the same word from text of the first audio segment and the second audio segment; (not explicitly disclosed)
plot the first audio segment and the second audio segment into corresponding first and second plots, wherein the first plot comprises a first curve comprising a first plurality of points, wherein the second plot comprises a second curve comprising a second plurality of points, (Pilz discloses inputting a speech sample. One of ordinary skill in the art would understand this to mean a waveform recording in a time-domain, which would include a plurality of points. [0032] discloses 2 speech samples: 1A and 1R which would represent first and second plots. )
and wherein a plot type of the first and second plots is a spectrogram; (Not explicitly disclosed by Pilz)
compare the first and second plots, wherein comparing comprises subtracting each point of the first plurality of points of the first curve from a corresponding point of the second plurality of points of the second curve at corresponding horizontal axis locations to form a difference plot; (Not explicitly disclosed by Pilz)
determine a quantity of outlier peaks, wherein the outlier peaks comprise peaks of the difference plot above an upper predetermined threshold or below a lower predetermined threshold; (Not explicitly disclosed by Pilz)
assign an artificial user probability to the user interaction, wherein the artificial user probability is low if the quantity of outlier peaks is greater than a predetermined outlier peak threshold; and ("[0032] … from the two current voice profiles Speaker Model 1A and Speaker Model 1R a current first similarity measure Measure 1 is obtained and subjected to a threshold-value discrimination DISCR1, with a predetermined upper threshold value. If this value is exceeded, the input speech samples are judged to have been fraudulently produced, and the attempted access is rejected. …" – exceeding the threshold indicates more similarity, which would be the same concept as staying under the threshold of outlier peaks.)
display the artificial user probability on a user interface of an endpoint device. (Fig. 1 shows "Similarity Measure 1" is calculated. Fig. 1, 101 is a server. One of ordinary skill in the art would understand that a server would connect to a display on an endpoint device which could display calculated parameters.)
Pilz does not disclose: spectrograms, determining similarity by subtracting plots and detecting outlier peaks, or identifying two instances of the same word from text.
SciPy discloses: determine a quantity of outlier peaks, ("Find peaks inside a signal based on peak properties." Para 1 )
wherein the outlier peaks comprise peaks of the difference plot above an upper predetermined threshold or below a lower predetermined threshold; ("height : number or ndarray or sequence, optional Required height of peaks. ..." This value defines a predetermined threshold for the peaks. Para 2)
Pilz and SciPy are considered analogous art to the claimed invention because they disclose methods of analyzing waveforms. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the fraud detection system of Pilz with the function disclosed by SciPy to count a number of outlier peaks as a similarity score, rather than calculating an average. This combination falls under combining prior art elements according to known methods to yield predictable results or simple substitution of one known element for another to obtain predictable results. See MPEP 2141, KSR, 550 U.S. at 418, 82 USPQ2d at 1396.
SciPy does not disclose: comparing spectrograms or identifying two instances of the same word from text.
Marius discloses: wherein a natural language processing engine identifies the first audio segment and the second audio segment as two instances of the same word from text of the first audio segment and the second audio segment (“an ASR algorithm converts an audio stream (or file) into a textual transcript of the spoken words. Detecting duplicates of the Maloney episodes can be implemented by applying ASR to an episode’s audio file, and comparing the resulting transcript with the transcript of all other, already transcribed episodes, whose texts are stored in a database.” Top of pg. 3)
Pilz, SciPy, and Marius are considered analogous art to the claimed invention because they disclose methods of analyzing waveforms. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the fraud detection system of Pilz in view of SciPy to detect two instances of the same word, as disclosed by Marius. This would have been beneficial because it is robust against noise (Marius pg. 4, top left box), and so that the user would not need to be prompted to repeat the word.
Marius further discloses: plot the first audio segment and the second audio segment into corresponding first and second plots, wherein the first plot comprises a first curve comprising a first plurality of points, wherein the second plot comprises a second curve comprising a second plurality of points, and wherein a plot type of the first and second plots a spectrogram; (Figure “Spectrogram peaks” on pg. 3 shows a spectrogram plot.)
compare the first and second plots, (see Acoustic Fingerprinting section on pg. 3.)
Pilz, SciPy, and Marius are considered analogous art to the claimed invention because they disclose methods of analyzing waveforms. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination to compare the audio as spectrograms as disclosed by Marius. This would have been beneficial for robustness, performance, accuracy, and lower set up and complexity. (Marius pg. 4, top right box)
Marius does not explicitly disclose that the spectrogram comparison comprises subtracting points on the curve.
James discloses: compare the first and second plots, wherein comparing comprises subtracting each point of the first plurality of points of the first curve from a corresponding point of the second plurality of points of the second curve at corresponding horizontal axis locations to form a difference plot; (“Substracting the two spectrograms on a point by point basis” pg. 1, line 4; see also figures on pgs. 2-3 which show subtraction of the spectrograms.)
Pilz, SciPy, Marius, and James are considered analogous art to the claimed invention because they disclose methods of analyzing waveforms. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination to compare the spectrograms as disclosed by James. This would have been beneficial to visualize the differences. (James pg. 1, 2nd para from end). This combination falls under combining prior art elements according to known methods to yield predictable results or simple substitution of one known element for another to obtain predictable results. See MPEP 2141, KSR, 550 U.S. at 418, 82 USPQ2d at 1396.
Regarding claim 2, Pilz discloses: 2. The system of claim 1, wherein the first audio segment is received subsequent a first word prompt provided to a user, wherein the second audio segment is provided in response to a second word prompt provided to a user, and wherein the first word prompt and the second word prompts are identical. ("[0032] ... For this purpose he is instructed to input a Speech Sample 1, for instance to speak into the system a password assigned to him, and from this currently acquired speech sample, again in a training process, a current voice profile Speaker Model 1A is produced. Immediately thereafter the alleged user is instructed to input the same speech sample again, and this new sample is available as Speech Sample 1R. ..." )
Regarding claim 4, Pilz discloses a system which records a password. Pilz does not explicitly disclose that the password contains multiple words, but it would be obvious to one of ordinary skill in the art that it could be a passphrase.
Marius discloses: 4. The system of claim 1, wherein the first audio segment and the second audio segment are captures of space between words, wherein the space between words is related to auditory cadence captured in the first and second audio segments. (Pg. 3, figure “Spectogram Peaks”, discloses performing the comparison for a song, which would include space between words related to auditory cadence. A spectrogram uses a time scale which would also apply to spoken words. The system of combination would be capable of capturing space between words which is related to auditory cadence.)
See claim 1 for motivation statement.
Regarding claim 5, Pilz discloses: 5. The system of claim 1, wherein the first audio segment and the second audio segment are captured in real-time from a singular audio source. ("[0032] ... For this purpose he is instructed to input a Speech Sample 1, for instance to speak into the system a password assigned to him, and from this currently acquired speech sample, again in a training process, a current voice profile Speaker Model 1A is produced. Immediately thereafter the alleged user is instructed to input the same speech sample again, and this new sample is available as Speech Sample 1R. ..." - Fig. 4 for example shows that the user's phone is the single audio source.)
Claim 8 is a computer program product claim with limitations corresponding to the limitations of Claim 1 and is rejected under similar rationale.
Claim 9 is a computer program product claim with limitations corresponding to the limitations of Claim 2 and is rejected under similar rationale.
Claim 11 is a computer program product claim with limitations corresponding to the limitations of Claim 4 and is rejected under similar rationale.
Claim 12 is a computer program product claim with limitations corresponding to the limitations of Claim 5 and is rejected under similar rationale.
Claim 15 is a method claim with limitations corresponding to the limitations of Claim 1 and is rejected under similar rationale.
Claim 16 is a method claim with limitations corresponding to the limitations of Claim 2 and is rejected under similar rationale.
Claim 18 is a method claim with limitations corresponding to the limitations of Claim 4 and is rejected under similar rationale.
Claim 19 is a method claim with limitations corresponding to the limitations of Claim 5 and is rejected under similar rationale.
Claim(s) 6-7, 13-14, and 20-21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Pilz in view of SciPy, Marius, and James as applied in claim 3 above, in further view of Wang et al. (US 20220358934 A1).
Regarding claim 6, Pilz, SciPy, Marius, and James do not disclose the additional limitations.
Wang discloses: 6. The system of claim 1, wherein if the artificial user probability is high, at least one of a spectrogram of the first audio segment and a spectrogram of the second audio segment is transmitted to a machine learning subsystem as training data. ("[0037] The evaluation unit evaluates the created multi-channel spectrogram by applying the generated multi-channel spectrogram to a classifier. The classifier is constructed using labeled multi-channel spectrograms as training data. The evaluation unit classifies the created multi-channel spectrogram to either genuine or spoof." Spectrograms of the audio segments along with the probability could be used as training data. Using all the audio segments as training data would include using the segments with high fraud probability.)
Pilz, SciPy, Marius, and James are considered analogous art to the claimed invention because they disclose methods of analyzing waveforms. Pilz and Wang are considered analogous art to the claimed invention because the disclose detecting fraudulent audio. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the combination with a machine learning subsystem as taught by Wang, and used the labeled data for training. Doing so would have been beneficial for higher detection accuracy. (Wang [0004])
Regarding claim 7, Pilz discloses: 7. The system of claim 6, wherein the at least one processing device is further configured to: query the machine learning subsystem, prior to comparing the first and second plots, to determine if the spectrogram is identical to a known spectrogram; (not explicitly disclosed)
and terminate the user interaction if the spectrogram is identical. ("[0032] ...If this value is exceeded, the input speech samples are judged to have been fraudulently produced, and the attempted access is rejected…." )
Pilz, SciPy, Marius, and James do not disclose querying a machine learning subsystem.
Wang discloses: 7. The system of claim 6, wherein the at least one processing device is further configured to: query the machine learning subsystem, prior to comparing the first and second plots, to determine if the spectrogram is identical to a known spectrogram; ("[0037] The evaluation unit classifies the created multi-channel spectrogram to either genuine or spoof." )
Rationale for combination as provided for Claim 6. Wang was cited for teaching the “machine learning subsystem” limitation of Claim 6 and Claim 7 provides further details regarding the same step which are combined under the same rationale.
Claim 13 is a computer program product claim with limitations corresponding to the limitations of Claim 6 and is rejected under similar rationale.
Claim 14 is a computer program product claim with limitations corresponding to the limitations of Claim 7 and is rejected under similar rationale.
Claim 20 is a method claim with limitations corresponding to the limitations of Claim 7 and is rejected under similar rationale.
Claim 21 is a method claim with limitations corresponding to the limitations of Claim 6 and is rejected under similar rationale.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Wang et al. (US,7627477 B2), hereinafter as Wang2. Wang2 discloses a method for comparing fingerprints of spectrogram plots. The method subtraction of points. (“Furthermore, if the audio sample is linearly stretched, such as simply being played back faster, then additionally frequency and delta time enjoy a reciprocal relationship, so that quantities such as f1*(t2−t1) are also invariant. Logarithms of these quantities may be used, substituting addition and subtraction for multiplication and division. To discover both the frequency and time stretch ratios, assuming they are independent, it is necessary to have both a frequency variant and a time variant quantity.” Col 4, last para)
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JON C MEIS whose telephone number is (703)756-1566. The examiner can normally be reached Monday - Thursday, 8:30 am - 5:30 pm EST.
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, Hai Phan can be reached at 571-272-6338. 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.
/JON CHRISTOPHER MEIS/Examiner, Art Unit 2654
/HAI PHAN/Supervisory Patent Examiner, Art Unit 2654