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
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 02/26/2025 and 07/17/2026 were filed and are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 103
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 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 1-6, 13-14, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Zainulina et al. (NO-REFERENCE DENOISING OF LOW-DOSE CT PROJECTIONS, hereinafter Zainulina) in view of QI et al. (US 2021/0290191 A1, hereinafter Qi).
Regarding Claim 1, Zainulina discloses
A method for training a machine-learning model for denoising, comprising: retrieving a target image data frame, the target image data frame being one image data frame of a sequence of image data frames containing imaging data of a subject (2.1 Noise2NoiseTD approach overview: “The proposed approach is based on the assumption that, given a sequence of 2k + 1image frames (CT-projections)… Thus, Bi-ConvLSTM units provide a stable restoration of the middle frame in the sequence that is being denoised.”);
retrieving at least one prior image data frame of the sequence of image data frames prior to the target image data frame in the sequence, wherein contents of the at least one prior image data frame overlap at least partially with the contents of target image data frame; retrieving at least one following image data frame of the sequence of image data frames following the target image data frame in the sequence, wherein contents of the at least one following image data frame overlap at least partially with contents of the target image data frame (2.1 Noise2NoiseTD approach overview: “The choice of the number of the adjacent frames k depends on how much the content of the frames overlaps, and the computational and memory capacity of the device…We propose to use the bidirectional convolutional memory units (Bi-ConvLSTM) for carrying information about the adjacent images. These units allow to extract the features corresponding to the slight consequent change of the structures that are observed from the first to the last viewing angle, and vice versa, and then combine these features.”);
generating a prediction for a denoised target image data frame based on the at least one prior image data frame and the at least one following image data frame (2.1 Noise2NoiseTD approach overview: “Thus, Bi-ConvLSTM units provide a stable restoration of the middle frame in the sequence that is being denoised…The task of this network is to process and fuse the results to obtain the denoised middle projection; therefore it can have a simpler architecture.”);
training a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters (2.2. Network architecture: “The output of the last Bi-ConvLSTM layer is summed up along the time axis, as we want to predict only the middle projection in the sequence. With the purpose of preventing over fitting to the noisy middle projection, we make the network “blind” to it by excluding the corresponding Bi-ConvLSTM output from the summation.”, 2.3 Loss function: “However, for low-dose CT the noise at each pixel can be more accurately modeled as an independent random variable sampled from a mixed Poisson-Gaussian distribution…where λ is the maximum event count and ɑ is the variance of the additive Gaussian noise.”, 3. EXPERIMENTS AND RESULTS: “The noise model (parameters ɑ and λ) was trained together with the main denoising model.”).
However, Zainulina does not explicitly disclose
retrieving acquisition parameters associated with the acquisition of the image data frames of the sequence of image data frames;
training a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters.
Qi teaches
retrieving acquisition parameters associated with the acquisition of the image data frames of the sequence of image data frames (Para [0033]: “In general, for X-ray CT, a count-domain projection
P
(
I
∙
T
)
) (where I denotes the tube current of the acquisition and T denotes the exposure time) can be approximated by a combination of a compound Poisson distributed transmission noise and a white Gaussian electronic noise”, Para [0033]-Para[0034]: “In an embodiment using half-dose scans, each half-scan is acquired at
t
2
current and exposure time T.”);
training a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters (Para [0003]: “a common approach is to decrease the tube current (Greess et al 2002). Unfortunately, lower dose introduces more noise and artifacts, which may deteriorate the diagnostic value of CT images.”, Para [0033]: “In general, for X-ray CT, a count-domain projection
P
(
I
∙
T
)
) (where I denotes the tube current of the acquisition and T denotes the exposure time) can be approximated by a combination of a compound Poisson distributed transmission noise and a white Gaussian electronic noise”, Para [0033]-Para[0034]: “In an embodiment using half-dose scans, each half-scan is acquired at
t
2
current and exposure time T.”).
It would be obvious to combine Zainulina and Qi before the effective filling date since Qi teaches that “In order to reduce the radiation dose, a common approach is to decrease the tube current” (Para [0003]) and that “lower dose introduces more noise and artifacts” (Para [0003]) in a projection where “where I denotes the tube current of the acquisition and T denotes the exposure time … can be approximated by a combination of a compound Poisson distributed transmission noise and a white Gaussian electronic noise” (Para [0033]). Where, “the mean and variance matching the Poisson plus Gaussian model of a true half-dose scan … can be used to train machine-based learning systems (e.g., artificial neural networks) for denoising low-dose CT images using a Noise2Noise training method” (Para [0040]). This is the same object of the instant application (Specification Para. 6). MPEP 2143(G).
Regarding Claim 2, dependent upon claim 1, Zainulina in view of Qi teaches everything regarding claim 1.
Zainulina further discloses
the prediction for the denoised target image data frame is an estimation of the mean and standard deviation of the denoised target image data frame based on a representation of at least one anatomical feature extracted from each of the at least one prior image data frame and the at least one following image data frame and wherein (2.1 Noise2NoiseTD approach overview: “Finally, the features extracted from the frames of the sequence, with the exception of the middle projection, are summed up and processed by another CNN.”, 2.3 Loss function: “Train the model to map the pixels in the patch into the mean
μ
x
and standard deviation
σ
x
of the Gaussian approximation of the distribution of the clean data
p
(
x
|
Ω
y
)
”)
the representations from the at least one prior image data frame and the at least one following image data frame are fused to form the prediction for the denoised target image data frame (2.1 Noise2NoiseTD approach overview: “Finally, the features extracted from the frames of the sequence, with the exception of the middle projection, are summed up and processed by another CNN.”).
Regarding Claim 3, dependent upon claim 2, Zainulina in view of Qi teaches everything regarding claim 2.
Zainulina further discloses
the representations of the at least one anatomical feature are transported between frames using convolutional memory units (2.1 Noise2NoiseTD approach overview: “We propose to use the bidirectional convolutional memory units (Bi-ConvLSTM) for carrying information about the adjacent images. These units allow to extract the features corresponding to the slight consequent change of the structures that are observed from the first to the last viewing angle, and vice versa, and then combine these features.”).
Regarding Claim 4, dependent upon claim 3, Zainulina in view of Qi teaches everything regarding claim 3.
Zainulina further discloses
the convolutional memory units are convolutional long short-term memory units for carrying information between frames of the sequence of image data frames (2.1 Noise2NoiseTD approach overview: “We propose to use the bidirectional convolutional memory units (Bi-ConvLSTM) for carrying information about the adjacent images. These units allow to extract the features corresponding to the slight consequent change of the structures that are observed from the first to the last viewing angle, and vice versa, and then combine these features.”).
Regarding Claim 5, dependent upon claim 2, Zainulina in view of Qi teaches everything regarding claim 2.
Zainulina further discloses
the at least one prior image data frame is a plurality of image data frames in the sequence of image data frames prior to the target image data frame and wherein the at least one following image data frame is a plurality of image data frames of the sequence of image data frames following the target image data frame (2.1 Noise2NoiseTD approach overview: “The choice of the number of the adjacent frames depends on how much the content of the k frames overlaps, and the computational and memory capacity of the device.”).
Regarding Claim 6, dependent upon claim 2, Zainulina in view of Qi teaches everything regarding claim 2.
Zainulina further discloses
the prediction for the denoised target image data frame is output by a trained convolutional neural network provided with the at least one prior image data frame and the at least one following image data frame (2.1 Noise2NoiseTD approach overview: “Finally, the features extracted from the frames of the sequence, with the exception of the middle projection, are summed up and processed by another CNN. The task of this network is to process and fuse the results to obtain the denoised middle projection; therefore it can have a simpler architecture.”).
Regarding Claim 13, dependent upon claim 1, Zainulina in view of Qi teaches everything regarding claim 1.
Zainulina further discloses
the machine learning algorithm is a convolutional neural network (2.1 Noise2NoiseTD approach overview: “We propose to use the bidirectional convolutional memory units (Bi-ConvLSTM) for carrying information about the adjacent images. These units allow to extract the features corresponding to the slight consequent change of the structures that are observed from the first to the last viewing angle, and vice versa, and then combine these features.”).
Regarding Claim 14, dependent upon claim 1, Zainulina in view of Qi teaches everything regarding claim 1.
Zainulina further discloses
the imaging data is CT imaging data, and wherein each image data frame of the sequence of image data frames is a projection frame, and wherein each projection frame comprises imaging data of the same subject acquired from a different angle (2.1 Noise2NoiseTD approach overview: “The proposed approach is based on the assumption that, given a sequence of image
2
k
+
1
frames (CT-projections),
p
θ
±
i
∆
θ
,
i
=
1
,
…
,
k
, where
θ
is the angle of rotation of the X-ray source around the patient, the content of the frames can be distinguished from the noise using similarities found among them by a neural network.”).
Regarding Claim 17, Zainulina discloses
A machine learning training system comprising: a memory that stores a plurality of instructions; and processor circuitry that couples to the memory and is configured to execute the instructions to: retrieve a plurality of image data frames comprising a sequence of image data frames containing imaging data of a subject; identify a target image data frame of the sequence of image data frames (2.1 Noise2NoiseTD approach overview: “The proposed approach is based on the assumption that, given a sequence of 2k + 1image frames (CT-projections)… Thus, Bi-ConvLSTM units provide a stable restoration of the middle frame in the sequence that is being denoised.”);
generate a prediction for a denoised target image data frame based on at least one prior image data frame of the sequence of image data frames prior to the target image data frame in the sequence and at least one following image data frame following the target image data frame in the sequence, wherein each of the at least one prior image data frame and the at least one following image data frame overlap at least partially with the target image data frame (2. 1 Noise2NoiseTD approach overview: “The choice of the number of the adjacent frames k depends on how much the content of the frames overlaps, and the computational and memory capacity of the device…We propose to use the bidirectional convolutional memory units (Bi-ConvLSTM) for carrying information about the adjacent images. These units allow to extract the features corresponding to the slight consequent change of the structures that are observed from the first to the last viewing angle, and vice versa, and then combine these features. Thus, Bi-ConvLSTM units provide a stable restoration of the middle frame in the sequence that is being denoised…The task of this network is to process and fuse the results to obtain the denoised middle projection; therefore it can have a simpler architecture”);
train a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters (2.2. Network architecture: “The output of the last Bi-ConvLSTM layer is summed up along the time axis, as we want to predict only the middle projection in the sequence. With the purpose of preventing over fitting to the noisy middle projection, we make the network “blind” to it by excluding the corresponding Bi-ConvLSTM output from the summation.”, 2.3 Loss function: “However, for low-dose CT the noise at each pixel can be more accurately modeled as an independent random variable sampled from a mixed Poisson-Gaussian distribution…where λ is the maximum event count and ɑ is the variance of the additive Gaussian noise.”, 3. EXPERIMENTS AND RESULTS: “The noise model (parameters ɑ and λ) was trained together with the main denoising model.”).
However, Zainulina does not explicitly disclose
a memory that stores a plurality of instructions; and processor circuitry that couples to the memory;
retrieve acquisition parameters associated with the acquisition of the image data frames of the sequence of image data frames;
train a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters.
Qi teaches
a memory that stores a plurality of instructions; and processor circuitry that couples to the memory (Fig. 15: “Memory 992, Processing circuitry 993”);
retrieve acquisition parameters associated with the acquisition of the image data frames of the sequence of image data frames (Para [0033]: “In general, for X-ray CT, a count-domain projection
P
(
I
∙
T
)
) (where I denotes the tube current of the acquisition and T denotes the exposure time) can be approximated by a combination of a compound Poisson distributed transmission noise and a white Gaussian electronic noise”, Para [0033]-Para[0034]: “In an embodiment using half-dose scans, each half-scan is acquired at
t
2
current and exposure time T.”);
train a machine-learning algorithm to denoise the target image data frame based on the prediction for the denoised target image data frame and a noise model based on the acquisition parameters (Para [0003]: “a common approach is to decrease the tube current (Greess et al 2002). Unfortunately, lower dose introduces more noise and artifacts, which may deteriorate the diagnostic value of CT images.”, Para [0033]: “In general, for X-ray CT, a count-domain projection
P
(
I
∙
T
)
) (where I denotes the tube current of the acquisition and T denotes the exposure time) can be approximated by a combination of a compound Poisson distributed transmission noise and a white Gaussian electronic noise”, Para [0033]-Para[0034]: “In an embodiment using half-dose scans, each half-scan is acquired at
t
2
current and exposure time T.”).
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Zainulina et al. (NO-REFERENCE DENOISING OF LOW-DOSE CT PROJECTIONS, hereinafter Zainulina) in view of QI et al. (US 2021/0290191 A1, hereinafter Qi) and Anonymous (DICOM PS3.6 2020d2020e - Data Dictionary, hereinafter DICOM Dictionary).
Regarding Claim 12, dependent upon claim 1, Zainulina in view of Qi teaches everything regarding claim 1.
However, Zainulina in view of Qi does not explicitly teach
the acquisition parameters are extracted from a DICOM file associated with the sequence of projection frames.
DICOM Dictionary teaches
the acquisition parameters are extracted from a DICOM file associated with the sequence of projection frames (P. 43:
(0018,1150)
Exposure Time
ExposureTime
IS
1
(0018,1151)
X-Ray Tube Current
XRayTubeCurrent
IS
1
).
It would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Zainulina in view of Qi with metadata of DICOM file of DICOM Dictionary, which is in the same field of endeavor of medical imaging, as DICOM is a standardized and commonly used medical image file type and would provide the necessary parameters for the system of Zainulina in view of Qi through DICOM’s metadata.
Allowable Subject Matter
Claims 7-10, 15-16, and 18-20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Relevant Prior Art Directed to State of Art
Zhou et al. (US 2019/0108904 A1, hereinafter Zhou) is prior art not applied in the rejection(s) above. Zhou discloses a medical image processing apparatus according to an embodiment comprises a memory and processing circuitry. The memory is configured to store a plurality of neural networks corresponding to a plurality of imaging target sites, respectively, the neural networks each including an input layer, an output layer, and an intermediate layer between the input layer and the output layer, and each generated through learning processing with multiple data sets acquired for the corresponding imaging target site. The processing circuitry is configured to process first data into second data using, among the neural networks, the neural network corresponding to the imaging target site for the first data, wherein the first data is input to the input layer, and the second data is output from the output layer.
LEE et al. (US 2020/0311490 A1, hereinafter Lee) is prior art not applied in the rejection(s) above. Lee discloses a method and apparatus to reduce the noise in medical imaging by training a deep learning (DL) network to select the optimal parameters for a convolution kernel of an adaptive filter that is applied in the data domain. For example, in X-ray computed tomography (CT) the adaptive filter applies smoothing to a sinogram, and the optimal amount of the smoothing and orientation of the kernel (e.g., a bivariate Gaussian) can be determined on a pixel-by-pixel basis by applying a noisy sinogram to the DL network, which outputs the parameters of the filter ( e.g., the orientation and variances of the Gaussian kernel). The DL network is trained using a training data set including target data (e.g., the gold standard) and input data. The input data can be sinograms generated by a low-dose CT scan, and the target data generated by a high-dose CT scan.
Xu et al. (Deformed2Self: Self-Supervised Denoising for Dynamic Medical Imaging, hereinafter Xu) is prior art not applied in the rejection(s) above. Xu discloses Deformed2Self, an end-to-end self-supervised deep learning framework for dynamic imaging denoising. It combines single-image and multi-image denoising to improve image quality and use a spatial transformer network to model motion between different slices.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOSHUA CHEN whose telephone number is (703)756-5394. The examiner can normally be reached M-Th: 9:30 am - 4:30pm ET F: 9:30 am - 2:30pm ET.
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, STEPHEN R KOZIOL can be reached at (408)918-7630. 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.
/J. C./Examiner, Art Unit 2665
/Stephen R Koziol/Supervisory Patent Examiner, Art Unit 2665
Prosecution Insights