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 .
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Response to Amendment
The amendment filed on September 9, 2026 has been entered.
The amendment of claims 1-9 has been acknowledged.
In view of the amendment, the 35 U.S.C. 112(b) rejections have been withdrawn.
Response to Arguments
Applicant’s arguments filed on September 9, 2026, with respect to the 35 U.S.C. 112(b) rejection of claim 3, have been fully considered and are persuasive. The rejection has been withdrawn.
Applicant’s arguments filed on September 9, 2026, with respect to the amended claims, have been fully considered but are moot because the arguments rely on newly added and/or amended claim limitations (e.g., “calibration coefficient … matches a noise variance”). The examiner has revised the rejections to match the new claim limitations
Claim Objections
Claim 5 is objected to because of the following informalities: Claim 5 recites the limitation “anatomical structures of the input image is preserved.” The limitation appears to contain a clerical/typographical error. The limitation has been interpreted as “anatomical structures of the input image are preserved” (due the “structures” being plural). Appropriate correction is required.
Claim Rejections - 35 USC § 103
Claim(s) 1-9 is/are rejected under 35 U.S.C. 103 as being unpatentable over Tang et al. (“CCN-CL: A content-noise complementary network with contrastive learning for low-dose computed tomography denoising,” Computers in Biology and Medicine 147 (2022) 105759), in view of Kim et al. (“Weakly-supervised progressive denoising with unpaired CT images,” Medical Image Analysis 71 (2021) 102065), hereinafter referred to as Tang and Kim, respectively.
Regarding claims 1 and 9, Tang teaches a low-dose computed tomography (LDCT) denoising neural network apparatus and method, the method performed in an LDCT denoising neural network apparatus, comprising:
a processor configured to:
provide an LDCT image as an input image to a convolutional neural network (CNN), and output a simulated normal-dose computed tomography (NDCT) image as an output image (Tang pg. 2 right column: “Suppose X denotes an LDCT image and Y represents the normal-dose CT (NDCT) image corresponding to X, then the relationship between X and Y can be described as X = Y + N where the N denotes the noise between the LDCT image and its corresponding NDCT image”; Tang pg. 3 left column: “our network takes an LDCT image as the input … and respectively output a predicted content and a predicted noise”; Tang Fig. 1: NDCT, Denoised, & LDCT); and
control the CNN to progressively remove the noise according to a dose level of a given LDCT image (Tang pg. 4 right column: “The Adam method [64] is selected to optimize the performance of our proposed network … the number of training epochs is set to 60 for all the networks” – the Adam method (Adaptive Moment Estimation) is an adaptive learning algorithm; Tang Fig. 1: the CNN performs denoising in each iteration).
However, Tang does not appear to explicitly teach performing noise-variance calibration on the input image fed to the CNN, by calculating a calibrated output image through a linear combination of the output image of the CNN and a discrepancy between the input image and the output image multiplied by a calibration coefficient, such that a noise variance of the calibrated output image matches a noise variance of the NDCT image.
Pertaining to the same field of endeavor, Kim teaches performing noise-variance calibration on the input image fed to the CNN, by calculating a calibrated output image through a linear combination of the output image of the CNN and a discrepancy between the input image and the output image multiplied by a calibration coefficient, such that a noise variance of the calibrated output image matches a noise variance of the NDCT image (Kim §3.2., pg. 3 & eq. (2): “The noise image is scaled by a factor of α and added to the original CT image to synthesize the noisier CT image z as follows: z = y + αñ. The value of α is a controllable parameter that determines the noise gap from z to y …if the ñ has the same noise variance as the inherent noise in y, the noise gap from z to y is α2/(1+α2). By regarding z as a target and y as an input, we aim to train our CNN denoiser to compensate the noise gap of α2/(1+α2)”; Kim 2021 Fig. 1: Input image, LDCT image & Output image).
Tang and Kim are considered to be analogous art because they are directed to neural network methods and systems for CT denoising. 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 content-noise complementary network with contrastive learning for LDCT denoising (as taught by Tang) to calibrate noise variations so that the noise variances match (as taught by Kim) because the combination increases the flexibility in data collection and allows to utilize any unpaired data at any dose levels (Kim Abstract).
Regarding claim 2, Tang, in view of Kim, teaches the apparatus of claim 1, wherein the processor is further configured to generate the LDCT image configured to have a varying dose levels by adding noise to an NDCT image (Tang pg. 2 right column discussed above, Tang Eq. (6) & pg. 2-3: “The other type aims at finding a mapping to predict the noise and then gain the content indirectly … we utilize content-noise complementary learning strategy”; also see Kim Fig. 1 & pg. 3 discussed above; Kim pg. 3 right column: “To better preserve image features such as sharp edges and the normal-dose noise texture”; Kim pg. 7 left column: “Our method was able to preserve the objects and the normal-dose noise texture while reducing the noise down to the level of NDCT”).
Regarding claim 3 , Tang, in view of Kim, teaches the apparatus of claim 2, wherein the processor is further configured to generate the varying dose levels by adjusting the intensity of the noise by controlling an alpha parameter of the NDCT image (Tang pg. 2-3, Eq. (6) & Kim Fig. 1, pg. 3 discussed above), and
wherein the alpha parameter is an adjustable parameter that determines a level of the noise variance to be added to the NDCT image, thereby generating the varying dose by producing LDCT images having different noise variances according to a value of the alpha parameter (Kim §3.2., pg. 3 & eq. (2) discussed above).
Regarding claim 4, Tang, in view of Kim, teaches the apparatus of claim 2, wherein the processor is further configured to train the CNN to minimize the discrepancy between the NDCT image and the simulated NDCT image (Tang pg. 4 left column: “We use the L1 loss to calculate the loss between the output and the corresponding NDCT image … The contrastive regularization loss of our network minimizes the gap between the output of network Xdenoised and the clear image Y while maximizing the gap from the input image X”).
Regarding claim 5, Tang, in view of Kim, teaches the apparatus of claim 1, wherein the processor is further configured to adjust adjusts the noise so that anatomical structures of the input image is preserved, based on the discrepancy between the output image and the input image (Tang Fig. 5 & pg. 5 left column: “the proposed CCN-CL network model performs better in tissue details preservation”; Kim Fig. 2).
Regarding claim 6, Tang, in view of Kim, teaches the apparatus of claim 5, wherein the processor is further configured to adjust the noise so that the output image has a noise variance higher than or equal to that of the NDCT image by a threshold (Kim §3.2., pg. 3, Fig. 1, & eq. (2) discussed above).
Regarding claim 7, Tang, in view of Kim, teaches the apparatus of claim 1, wherein the processor is further configured to recognize the dose level of the given LDCT image through the CNN and progressively remove the noise by performing repeated input/output operations through the CNN (Tang Fig. 1 & pg. 4 right column discussed above; also see Kim pg. 5 right column: “Adam optimizer (Kingma and Ba, 2014) was used with the learning rate of 10-4 and default values for other parameters”).
Regarding claim 8, Tang, in view of Kim, teaches the apparatus of claim 7, wherein the processor is further configured to input an N-th (where N is a natural number) intermediate image to the CNN to generate an (N+1)-th intermediate image and again inputs the (N+1)-th intermediate image to the CNN to generate an (N+2)-th intermediate image (Tang Fig. 1, pg. 4 right column & Kim pg. 3 discussed above; Kim Figs. 10-11 & pg. 8 right column: “intermediate output images as well as the final output images”).
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 SOO J SHIN whose telephone number is (571)272-9753. The examiner can normally be reached M-F; 10-6.
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, Matthew Bella can be reached at (571)272-7778. 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.
/Soo Shin/Primary Examiner, Art Unit 2667 571-272-9753
soo.shin@uspto.gov
