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 .
Claims Accounting
Applicant's arguments, filed 02/26/2026, have been fully considered.
The following rejections are either reiterated or newly applied. They constitute the complete set presently being applied to the instant application.
Applicants have amended their claims, filed 02/26/2026, and therefore rejections newly made in the instant office action have been necessitated by amendment.
Claims 1, 4-5, and 7 have been amended.
Claims 2-3 and 8 have been canceled.
Claims 9 and 10 are newly presented.
Claims 1, 4-7, and 9-10 are the current claims hereby under examination.
Claim Objections
Claims 9 and 10 objected to because of the following informalities:
Claim 9 recites “an image having a plurality of pixel corresponding to the spectral reflectance of the skin imaga plurality of pixel corresponding to the spectral reflectance of the skin imaged by the camera” in lines 5-7. This should read “an image having a plurality of pixels corresponding to the spectral reflectance of the skin imaged by the camera”.
Claim 10 recites “an image having a plurality of pixel corresponding to the spectral reflectance of the skin imaga plurality of pixel corresponding to the spectral reflectance of the skin imaged by the camera” in lines 4-6. This should read “an image having a plurality of pixels corresponding to the spectral reflectance of the skin imaged by the camera”.
Appropriate correction is required.
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.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 4 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claim 4, the claim recites “wherein the loss function is a function for calculating the discrepancy for each wavelength when the image is visualized” in lines 5-6. It is unclear what is meant by “when the image is visualized”. There are no limitations drawn to the visualization of the image, and therefore it is unclear when the image is realized or what effect the visualization of the image may have in relation to the other functional limitations. Clarification is requested.
For the purposes of examination, the claim is interpreted as “wherein the loss function is a function for calculating the discrepancy for each wavelength of the image”.
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, 4-7, and 9-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim(s) as a whole, considering all claim elements both individually and in combination, do not amount to significantly more than an abstract idea. A streamlined analysis of claim 1 follows.
Step 1
Regarding claim 1, the claim recites a camera and processing circuitry configured to complete series of functional limitations, including to convert observation data to a feature quantity of the target event using a machine learning model, to restore observation data from the feature quantity using a numerical simulation model, and to train a machine learning model. Thus, the claim is directed to a machine, which is one of the statutory categories of invention.
Step 2A, Prong One
The claim is then analyzed to determine whether it is directed to any judicial exception. The functional limitations of converting observation data to a feature quantity of the target event using a machine learning model, restoring observation data from the feature quantity using a numerical simulation model, and training the machine learning model based on a loss function wherein the loss function is a function to calculate a discrepancy for each of the plurality of pixels set forth judicial exceptions. These functional limitations describe sets of mathematical calculations using formulas and/or equations. Thus, the claim is drawn to a Mathematical Concept, which is an Abstract Idea.
Step 2A, Prong Two
Next, the claim as a whole is analyzed to determine whether the claim recites additional elements that integrate the judicial exception into a practical application. The claim fails to recite an additional element or a combination of additional elements to apply, rely on, or use the judicial exception in a manner that imposes a meaningful limitation on the judicial exception. Claim 1 recites training the machine learning model, which is a judicial exception, as it is based on a loss function that calculates a discrepancy. Further, the training of the machine learning does not provide an improvement to the technological field, the method does not effect a particular treatment or effect a particular change based on the trained machine learning model, nor does the method use a particular machine to perform the Abstract Idea. It is noted that the processing circuitry is a generic computer component configured to perform the Abstract Idea. According to section 2106.05(f) of the MPEP, merely using a computer as a tool to perform an abstract idea does not integrate the Abstract Idea into a practical application.
Step 2B
Next, the claim as a whole is analyzed to determine whether any element, or combination of elements, is sufficient to ensure that the claim amounts to significantly more than the exception. Besides the Abstract Ideas, the claim recites additional functional limitations of acquiring observational data. The acquiring limitation is recited at a high level of generality such that it amounts to insignificant presolution activity, e.g., mere data gathering step necessary to perform the Abstract Idea. When recited at this high level of generality, there is no meaningful limitation, such as a particular or unconventional step that distinguishes it from well-understood, routine, and conventional data gathering and comparing activity engaged in by medical professionals prior to Applicant's invention. The claim also recites the limitation of a camera configured image a spectral reflectance of a skin of a patient, which, as recited, is a generic camera capable of collecting spectral images used to complete the data gathering acquiring step. Furthermore, it is well established that the mere physical or tangible nature of additional elements such as the processing circuitry acquiring an steps do not automatically confer eligibility on a claim directed to an abstract idea (see, e.g., Alice Corp. v. CLS Bank Int'l, 134 S.Ct. 2347, 2358-59 (2014)). It is noted that the limitation related to the feature quantity are details of the operation of the machine learning model, which is identified as an Abstract Idea.
The inclusion of the camera with the processing circuitry does not constitute a particular machine, as it comprises a generic camera (i.e., optical sensor) configured to perform the pre-solution data gathering step and a generic computer system configured to perform the judicial exception. According to section 2106.05(f) of the MPEP, merely using a computer as a tool to perform an abstract idea does not integrate the Abstract Idea into a practical application.
Consideration of the additional elements as a combination also adds no other meaningful limitations to the exception not already present when the elements are considered separately. Unlike the eligible claim in Diehr in which the elements limiting the exception are individually conventional, but taken together act in concert to improve a technical field, the claim here does not provide an improvement to the technical field. Even when viewed as a combination, the additional elements fail to transform the exception into a patent-eligible application of that exception. Thus, the claim as a whole does not amount to significantly more than the exception itself. The claim is therefore drawn to non-statutory subject matter.
The same rationale applies to claims 7 and 9-10.
The dependent claims also fail to add something more to the abstract independent claims as they generally recite functional limitations pertaining to the Abstract Ideas. Claims 4-5 are directed towards details of the loss function and claim 5 is directed towards details of the numerical simulation, both of which are identified as Abstract Ideas above. The converting, restoring, and training steps recited in the independent claims maintain a high level of generality even when considered in combination with the dependent claims.
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, 4-7, and 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over US Patent Publication 2020/0193597 by Fan et al. – previously cited, hereinafter “Fan”, in view of US Patent Publication 2021/0193298 by Sun et al. – previously cited, hereinafter “Sun”, as evidenced by US Patent Publication 2009/0317856 by Mycek et al. – previously cited, hereinafter “Mycek”, in view of US Patent Publication 2022/0011224 by Bertsimas et al. – previously cited, hereinafter “Bertsimas”.
Regarding claim 1, Fan teaches an information processing device comprising: a camera configured to image a spectral reflectance of a skin of a patient ([0160-0161]; healing predictions may be accomplished based on one or more images of the wound captured using a multispectral image sensor (i.e., camera)); and processing circuitry configured to ([0234]; All methods and tasks may be performed by a computer system and may be implemented in application-specific circuitry.): acquire observation data, which is acquired when a target event is observed and includes an image having a plurality of pixels corresponding to the spectral reflectance of the skin imaged by the camera (Fig. 16, multispectral image sensor as described in par. [0161] is used to acquire the ulcer image. Figs. 24 and 28; Input of ulcer image to the autoencoder, the target event is considered when the capturing of the observation data for training. [0188-0190]; The image data contains 3 channels per pixel, the 3 channels represented the diffuse reflectance of light from the tissue at the wavelengths from the skin, where an ulcer would occur as shown in Fig. 16); convert the observation data to a feature quantity of the target event using a machine learning model (Fig. 24, [0161]; The encoder of the autoencoder can “generate a reduced feature representation of the input, here a reduced number of values (e.g., numerical values) representing the pixel values in the input image(s)”); restore expected observation data from the feature quantity ([0169]; the decoder of the autoencoder: “then decompress this encoded data with the decoder such that the output is a good/perfect reconstruction of the original input data”); and train the machine learning model ([0167]; Neural networks are trained through backpropagation in which the network parameters are tuned to produce expected outputs given corresponding inputs in labeled training data.) based on a loss function based on a discrepancy between first observation data, which is the observation data that has not been converted to the feature quantity, and second observation data, which is the expected observation data restored from the feature quantity ([0190]; “For each predicted pixel value (in R3) of the output of the decoder layer, the loss was computed with mean square error (MSE) wherein the target values were the pixel values of the original image”).
Fan does not teach restoring the observation data from the feature quantity by using a numerical simulation model.
Fig. 3 of Sun teaches a method including a step (Step S206) of generating (i.e., restoring) a sample training image using feature data generated from the original training image from a neural network model (i.e., machine learning model) ([0039, 0051]). The training image is generated using a generative model, which may be a numerical simulation using a Monte-Carlo sampling method ([0020]).
It would have been prima facie obvious to one of ordinary skill in the art at the time of the effective filing date to have modified the device of Fan such that the processing circuitry is configured to use a numerical simulation model to restore the observation data from the feature quantity, as taught by Sun ([0020, 0039, 0051]). Mycek teaches that a Monte Carlo simulation is a computational method that is accurate throughout the optical parameter space for modeling photon transport in biological tissue ([0080]), and therefore the combination of Fan and Sun would use an accurate computational method.
Fan in view of Sun does not teach the loss function being based on the assumption that a frequency distribution of the discrepancy conforms to a predetermined probability density distribution, or wherein the loss function is a function to calculate the discrepancy for each of the plurality of pixels included in the image, the function being based on the assumption that the frequency distribution of the discrepancy for each of the plurality of pixels conforms to the predetermined probability density distribution.
Sun teaches a method of training a machine learning model. Sun teaches determining one or more training distribution divergences based at least in part on comparing the one or more training distributions to a reference distribution (i.e., predetermined probability density function), such as a normal distribution having a mean of zero and a variance of unity. A loss corresponding to the one or more training distribution divergences is determined ([0027]).
It would have been prima facie obvious to one of ordinary skill in the art at the time of the effective filing date to have modified the processing circuitry of Fan in view of Sun such that the machine learning model is trained based on a loss being based on the assumption that a frequency distribution of the discrepancy conforms to a predetermined probability density distribution, as taught by Sun ([0027]). This modification merely comprises a simple substitution of one known element (loss based on a frequency distribution conforming to a predetermined probability density distribution) for another (loss based on mean square error of target values vs original values) to obtain predictable results. See MPEP 2143.I.B. It is noted that Fan uses the loss function to train the machine learning model such that the loss is minimized, therefore, in the combination of Fan and Sun, the distribution divergences of would be minimized. It is noted that the method of Fan is directed towards the analysis of each pixel value in the image, and therefore the loss function would be calculated for each of the pixels. Therefore, the combination of Fan and Sun would result in wherein the loss function is a function to calculate the discrepancy for each of the plurality of pixels included in the image, the function being based on the assumption that the frequency distribution of the discrepancy for each of the plurality of pixels conforms to the predetermined probability density distribution.
Fan in view of Sun does not teach wherein the feature quantity is an in-vivo component volume of the patient.
Fig. 1B of Bertsimas teaches a method wherein the encoder of an autoencoder generates feature values (i.e., latent variables) in step 106B ([0069]). Bertsimas teaches that conventional techniques of dimension reduction provide a set of latent variables that may not provide a human interpretable indication of characteristics of a biological sample (i.e., in-vivo component volume). The feature generation step of Bertsimas identifies a subset of wavelengths of the spectral data that indicate characteristics of a biological sample, allowing a clinician to interpret a diagnosis result ([0051]).
It would have been prima facie obvious to one of ordinary skill in the art at the time of the effective filing date to have modified the processing circuitry of Fan in view of Sun such that the feature quantity is an in-vivo component volume of the patient, to provide a human interpretable indication of characteristics of a biological sample that allows a clinician to interpret a diagnosis result, as taught by Bertsimas ([0051]).
Regarding claim 4, the combination of Fan, Sun, and Bertsimas teaches the information processing device according to claim 1, and wherein the loss function is a function for calculating the discrepancy for each wavelength of the spectral reflectance when the image is visualized (Fan, [0188-0190]; The loss function is calculated for the data of each pixel, and each pixel has 3 channels representing the diffuse reflectance of light from the tissue at the wavelengths. Therefore, the loss function calculates the discrepancy for each wavelength), the function being based on the assumption that the frequency distribution of the discrepancy for each wavelength conforms to the predetermined probability density distribution (See the rejection of claim 1).
Regarding claim 5, the combination of Fan, Sun, and Bertsimas teaches the information processing device according to claim 1, wherein the loss function is a function for outputting a smaller loss as the discrepancy becomes less and as a distance between the frequency distribution of the discrepancy and the predetermined probability density distribution becomes less (Sun, [0024]; The loss function relied upon in the combination is that of Sun. The second loss corresponds to the distribution divergence, which is based on a comparison of the one or more attribute distributions to a reference distribution. The second loss being determined based on a divergence based on a comparison indicates that when a distance (or divergence) is less, the second loss also decreases.), wherein the processing circuitry is further configured to train the machine learning model such that the loss decreases (Sun, [0027]; “the training module 30 is configured to change… one or more parameters of the recognition model… to reduce, such as minimize the first loss and/or the second loss”).
Regarding claim 6, the combination of Fan, Sun, and Bertsimas teaches the information processing device according to claim 1, wherein the machine learning model is a model using a neural network or a genetic algorithm (Fan, [0161]; the machine learning model is an autoencoder, which is a type of neural network), and wherein the numerical simulation model is a model using a Monte Carlo method (Sun, [0020]; The numerical method relied upon is that of Sun. See the rejection of claim 1. The sample image is generated by the generative model by sampling the one or more attribute distributions corresponding to the medical image. The sampling method can be a Monte-Carlo sampling method), the Kubelka-Munk theory, or the Lambert‐Beer law.
Regarding claim 7, the combination of Fan, Sun, and Bertsimas teaches an information processing method performed by a computer (Fan, [0234]; All methods and tasks may be performed by a computer system and may be implemented in application-specific circuitry.), the information processing method comprising: acquiring observation data, which is acquired when a target event is observed and includes an image having a plurality of pixels corresponding to a spectral reflectance of a skin of a patient, the image being imaged by a camera (See the rejection of claim 1); converting the observation data to a feature quantity of the target event using a machine learning model (See the rejection of claim 1); restoring expected observation data from the feature quantity using a numerical simulation model (See the rejection of claim 1); and training the machine learning model based on an assumption that a frequency distribution of a discrepancy conforms to a predetermined probability distribution, the discrepancy being a discrepancy between first observation data, which is the observation data that has not been converted to the feature quantity, and second observation data, which is the expected observation data restored from the feature quantity (See the rejection of claim 1), wherein the feature quantity is an in-vivo component volume of the patient (See the rejection of claim 1), and wherein the loss function is a function to calculate the discrepancy for each of the plurality of pixels included in the image, the function being based on the assumption that the frequency distribution of the discrepancy for each of the plurality of pixels conforms to the predetermined probability density distribution (See the rejection of claim 1).
Regarding claim 9, the combination of Fan, Sun, and Bertsimas teaches an information processing device comprising: a camera configured to image a spectral reflectance of a skin of a patient (See the rejection of claim 1); processing circuitry configured (See the rejection of claim 1) to: acquire observation data, which is acquired when a target event is observed and includes an image having a plurality of pixels corresponding to a spectral reflectance of a skin of a patient, the image being imaged by a camera (See the rejection of claim 1); convert the observation data to a feature quantity of the target event using a machine learning model (See the rejection of claim 1); restore expected observation data from the feature quantity using a numerical simulation model (See the rejection of claim 1); and train the machine learning model based on an assumption that a frequency distribution of a discrepancy conforms to a predetermined probability distribution, the discrepancy being a discrepancy between first observation data, which is the observation data that has not been converted to the feature quantity, and second observation data, which is the expected observation data restored from the feature quantity (See the rejection of claim 1), wherein the feature quantity is an in-vivo component volume of the patient (See the rejection of claim 1), and wherein the loss function is a function to calculate the discrepancy for each wavelength of the spectral reflectance (See the rejection of claim 4), the function being based on the assumption that the frequency distribution of the discrepancy for each wavelength conforms to the predetermined probability density distribution (See the rejection of claim 4).
Regarding claim 10, the combination of Fan, Sun, and Bertsimas teaches an information processing method performed by a computer (Fan, [0234]; All methods and tasks may be performed by a computer system and may be implemented in application-specific circuitry.), the information processing method comprising: acquiring observation data, which is acquired when a target event is observed and includes an image having a plurality of pixels corresponding to a spectral reflectance of a skin of a patient, the image being imaged by a camera (See the rejection of claim 1); converting the observation data to a feature quantity of the target event using a machine learning model (See the rejection of claim 1); restoring expected observation data from the feature quantity using a numerical simulation model (See the rejection of claim 1); and training the machine learning model based on an assumption that a frequency distribution of a discrepancy conforms to a predetermined probability distribution, the discrepancy being a discrepancy between first observation data, which is the observation data that has not been converted to the feature quantity, and second observation data, which is the expected observation data restored from the feature quantity (See the rejection of claim 1), wherein the feature quantity is an in-vivo component volume of the patient (See the rejection of claim 1), and wherein the loss function is a function to calculate the discrepancy for each wavelength of the spectral reflectance (See the rejection of claim 4), the function being based on the assumption that the frequency distribution of the discrepancy for each wavelength conforms to the predetermined probability density distribution (See the rejection of claim 4).
Response to Arguments
Applicant’s arguments, filed 02/26/2026 have been fully considered.
The amendments to the claims overcome the rejections under 35 U.S.C. 112(b) of claims 1 and 7. The amendments to the claims do not overcome the rejection under 35 U.S.C. 112(b) of claim 4.
Applicant’s arguments regarding the rejections of the claims under 35 U.S.C. 101 are acknowledged. These arguments are not found persuasive. Applicant’s arguments regarding the training step not reciting a judicial exception is not found persuasive. Claims 1, 7, 9, and 10 recite the training step being based on a loss function, wherein the loss function is a function to calculate the discrepancy for each of the plurality of pixels in the image. The recitation of calculating the discrepancy is a clear recitation of a Mathematical Concept (i.e., mathematical calculation) during the training step, and is therefore drawn to a judicial exception. Applicant’s arguments that the recitation of a camera to generate an image having a plurality of pixels corresponding to the spectral reflectance of skin a patient integrates the Abstract idea into practical application is not found persuasive. The camera is a generic camera configured to obtain the image recited in the acquiring step, and is therefore drawn to a generic device configured to perform the pre-solutional activity of data gathering for the Abstract Idea.
Applicant’s assertion regarding the rejection of the independent claims under 35 U.S.C. 102 is acknowledged. This assertion is moot as it is based on amendments to the claims not entered at the time of the previous Office action. The newly presented limitations are rejected on new grounds above.
Applicant’s arguments regarding the rejections of the independent claims under 35 U.S.C. 103, that the each of the references fail to disclose the processing circuitry configured to train the machine learning model limitations are acknowledged. These arguments are not found persuasive. The arguments are directed toward why each reference do not disclose all of the limitations related to training the machine learning model. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). The rejections under 35 U.S.C. 103 in this Office action and the previous Office action rely on the combination of references.
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 reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NELSON A GLOVER whose telephone number is (571)270-0971. The examiner can normally be reached Mon-Fri 8:00-5:00 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, Jason Sims can be reached at 571-272-7540. 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.
/NELSON ALEXANDER GLOVER/Examiner, Art Unit 3791
/ADAM J EISEMAN/Primary Examiner, Art Unit 3791
Prosecution Insights