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 .
Claim Rejections - 35 USC § 102
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-2 and 13-17 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Amemiya et al (US 20230142011 A1), hereinafter Amemiya.
-Regarding claim 1, Amemiya discloses a method for reconstructing denoised magnetic resonance images, the method comprising (Abstract; FIGS. 1-14): (a) accessing k-space data with a computer system, wherein the k-space data have been acquired using a multichannel receiver (FIG. 1, coil 106, receiver 108; FIG. 2, data reception unit 231; [0035], “reception coil includes a plurality of channels …”; [0037], “… a position and a patter in a k-space …”; [0056]); (b) reconstructing coil channel images from the k-space data using the computer system, wherein each coil channel image corresponds to a different channel of the multichannel receiver (FIGS. 1-4, 9-14; [0010]; [0035], “each channel … each of the reception coils”; [0037], “The measurement data for each channel corresponding to the small-sized reception coils is arranged in a k-space …”; [0089], “creates a reconstructed image … for each channel …”); (c) generating denoised coil channel images with the computer system by applying a denoising algorithm to the coil channel images using the computer system (FIG. 2, noise removal unit 237; [0067], “denoising is performed by using a predetermined algorithm …”; FIG. 3, step S303; FIGS. 4, 7, 9-10; FIGS. 11-14); (d) generating denoised k-space data from the denoised coil channel images using the computer system to transform the denoised coil channel images into k-space (FIG. 2, unit 238; FIG. 3, S304, S3042; FIGS. 4, 9-12, 14; [0013], “… creating denoised k-space data for each of the reception coils by using a denoised image …”); and (e) reconstructing denoised magnetic resonance images from the denoised k-space data using the computer system (FIGS. 2-4, 9-14) by applying the denoised k-space data to a nonlinear reconstruction algorithm (FIG. 7; [0023]; [0068], “a deep learning model such as a convolutional neural network (CNN) learned to remove Gaussian noise”; [0070]; FIG. 8), generating output as the denoised magnetic resonance images (FIGS, 2-3; FIG. 4, reconstructed image; FIGS. 9-14).
-Regarding claim 2, Amemiya discloses the method of claim 1. Amemiya further discloses wherein reconstructing the denoised magnetic resonance images comprises: accessing a neural network with the computer system, wherein the neural network has been trained on training data to reconstruct magnetic resonance images from k-space data based on a nonlinear image reconstruction framework; and applying the denoised k-space data to the neural network, generating output as the denoised magnetic resonance images (FIGS. 2-4, 7-8; [0010], “estimated measurement data by integrating the denoised k-space data … The iterative reconstruction unit … based on the estimated measurement data”; [0070], “The output of the CNN … a denoised image …”; [0074]; [0084]; [0103], “a plurality of CNNs, such as a CNN learned with training data having a large noise amount and a CNN learned with training data having a small noise amount …”).
-Regarding claim 13, Amemiya discloses the method of claim 1. Amemiya further discloses wherein the denoising algorithm is a channel-independent denoising algorithm ([0047], “the image processing unit 230 repeatedly performs … image synthesis for each channel, removal of noise included in the images, consistency maintenance, and the like”; [0087], “denoising is performed on an image for each channel”; [0089]).
-Regarding claim 14, Amemiya discloses the method of claim 1. Amemiya further discloses wherein the denoising algorithm comprises a neural network that has been trained on training data to denoise an input image, wherein generating the denoised coil channel images with the computer system comprises inputting the coil channel imaged to the neural network, generating the denoised coil channel images as an output (FIGS. 2-4, 7-8; [0103]).
-Regarding claim 15, Amemiya discloses the method of claim 14. Amemiya further discloses wherein the neural network is a convolutional neural network. (FIGS. 7-8).
-Regarding claim 16, Amemiya discloses the method of claim 1. Amemiya further discloses wherein the k-space data accessed with the computer system comprise undersampled k-space data (FIG. 1, 5A-5D; [0011]; [0044]).
-Regarding claim 17, Amemiya discloses the method of claim 1. Amemiya further discloses wherein the coil channel images comprise a dynamic series of images (FIGS. 2-4, 9-10; [0044], “iterative reconstruction …”; [0061]; [0066]).
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.
Claim(s) 3-7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Amemiya et al (US 20230142011 A1), hereinafter Amemiya in view of Yaman et al (arXiv:2008.06029v1 2020), hereinafter Yaman.
-Regarding claim 3, Amemiya discloses the method of claim 2.
Amemiya does not disclose wherein the nonlinear image reconstruction framework includes a physics-guided deep learning image reconstruction.
In the same field of endeavor, Yaman teaches a self-supervised learning method for physics-guided neural networks in highly accelerated MRI (Yaman: Abstract; FIGS. 1-6, S1-S4). Yaman further teaches wherein the nonlinear image reconstruction framework includes a physics-guided deep learning image reconstruction (Yaman: Abstract; FIG. 1; Page 6, Sec., Methods).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Yaman by using the nonlinear image reconstruction framework that includes a physics-guided deep learning image reconstruction in order to address issues such as aliasing artifact when data is scarce or when predictions must obey known physical laws by embedding physical knowledge into the learning process.
-Regarding claim 4, Amemiya discloses the method of claim 2.
Amemiya does not disclose wherein the neural network has been trained on training data using self-supervised learning.
In the same field of endeavor, Yaman teaches a self-supervised learning method for physics-guided neural networks in highly accelerated MRI (Yaman: Abstract; FIGS. 1-6, S1-S4). Yaman further teaches wherein the neural network has been trained on training data using self-supervised learning (Yaman: Abstract; FIG. 1; Page 7, 2nd paragraph; Page 14, Sec. Discussion).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Yaman by using self-supervised learning method in order to reduce the reliance on labeled data.
-Regarding claim 5, Amemiya in view of Yaman teaches the method of claim 4. The combination further teaches wherein the neural network has been trained on training data by separating the training data into a first subset of training data and a second subset of training data, wherein the first subset of training data is used within the neural network during training and the second subset of training data is used in a loss function used during training (Yaman: FIG. 1, caption “The acquired sub-sampled k-space measurements for each scan, Ω, are split into multiple disjoint sets, Θ𝑗 and Λ𝑗 in which Ω = Θ𝑗 ∪ Λ𝑗 , for j ∈ {1, …, K}. The first set of indices, Θ𝑗 , is used in the data consistency unit of the unrolled network, while the latter set, Λ𝑗 is used to define the loss function for training …”).
-Regarding claim 6, Amemiya in view of Yaman teaches the method of claim 5. The combination further teaches wherein the first subset of training data defines data consistency units and the second subset of training data defined k-space loss (Yaman: FIG. 1, caption “The acquired sub-sampled k-space measurements for each scan, Ω, are split into multiple disjoint sets, Θ𝑗 and Λ𝑗 in which Ω = Θ𝑗 ∪ Λ𝑗 , for j ∈ {1, …, K}. The first set of indices, Θ𝑗 , is used in the data consistency unit of the unrolled network, while the latter set, Λ𝑗 is used to define the loss function for training …”).
-Regarding claim 7, Amemiya discloses the method of claim 2. Amemiya further discloses wherein the k-space data have been acquired from a subject (FIG. 1, subject 101).
Amemiya does not disclose wherein the training data comprise subject-specific training data also acquired from the subject.
In the same field of endeavor, Yaman teaches a self-supervised learning method for physics-guided neural networks in highly accelerated MRI (Yaman: Abstract; FIGS. 1-6, S1-S4). Yaman further teaches wherein the k-space data have been acquired from a subject and the training data comprise subject-specific training data also acquired from the subject (Yaman: Abstract; FIG. 1; Page 7, 1st paragraph, “same subject i”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Yaman by using the nonlinear image reconstruction framework that includes a physics-guided deep learning image reconstruction in order to address issues such as aliasing artifact when data is scarce or when predictions must obey known physical laws by embedding physical knowledge into the learning process.
Claim(s) 8-12 and 18-25 is/are rejected under 35 U.S.C. 103 as being unpatentable over Amemiya et al (US 20230142011 A1), hereinafter Amemiya in view of Moeller et al (US 20190086496 A1), hereinafter Moeller.
-Regarding claim 8, Amemiya discloses the method of claim 1.
Amemiya does not disclose wherein the denoising algorithm comprises a locally low- rank (LLR)-based denoising algorithm.
In the same field of endeavor, Moeller teaches a method for controlling noise in magnetic resonance imaging (Moeller: Abstract; FIGS. 1-5). Moeller further teaches wherein the denoising algorithm comprises a locally low-rank (LLR)-based denoising algorithm (Moeller: FIG. 3A; [0020], “A local low rank denoising technique may then be applied … using the matrix and the series of images”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Moeller by using LLR in order to model correlation between images that acquired at different contrasts, and improve fidelity, especially when only a few images are available.
-Regarding claim 9, Amemiya in view of Moeller teaches the method of claim 8. The combination further teaches wherein the LLR-based denoising algorithm comprises: selecting, with the computer system, an image patch corresponding to the coil channel images; forming a matrix with the computer system by combining vectors generated using the image patch; and applying a locally low-rank denoising with the computer system using the matrix and the coil channel images to generate the denoised coil channel images (Moeller: FIG. 1; [0008], “An image patch corresponding to the series of images is selected with the computer system, and a matrix is formed with the computer system by combining vectors generated using the image patch. A local low-rank denoising technique is applied with the computer system using the matrix and the series of images to generate at least one denoised image”; [0027]).
-Regarding claim 10, Amemiya in view of Moeller teaches the method of claim 9. The combination further teaches wherein the locally low-rank denoising implements a singular value decomposition (Moeller: FIG. 3A; [0028]).
-Regarding claim 11, Amemiya in view of Moeller teaches the method of claim 10. The combination further teaches wherein the locally low-rank denoising implements a singular value thresholding (Moeller: FIG. 3A; [0027], “enforced via singular value thresholding”; [0028]).
-Regarding claim 12, Amemiya in view of Moeller teaches the method of claim 11. The combination further teaches wherein the singular value thresholding is implemented using a threshold value computed based on singular values of a random Gaussian matrix (Moeller: [0028], “determined from the singular values of a random Gaussian matrix …”).
-Regarding claim 18, Amemiya discloses the method of claim 1.
Amemiya does not disclose wherein the coil channel images comprise a contrast- varying series of images.
In the same field of endeavor, Moeller teaches a method for controlling noise in magnetic resonance imaging (Moeller: Abstract; FIGS. 1-5). Moeller further teaches wherein the coil channel images comprise a contrast- varying series of images (Moeller: FIG. 4, acquisition server 412; [0027], “… redundancies in the image series that acquire the same anatomy over time or through contrast changes (e.g., across different contrast weightings) …”; [0043], “contrast agent in a magnetic resonance angiography (“MRA”) scan”).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Moeller by using coil channel images that comprise a contrast-varying series of images in order to model correlation between images that acquired at different contrasts, and reduce noise while keeping subtle variations intact.
-Regarding claim 19, Amemiya discloses a method for reconstructing denoised magnetic resonance images, the method comprising (Abstract; FIGS. 1-14): (a) accessing undersampled k-space data with a computer system, wherein the undersampled k-space data have been acquired using a multichannel receiver (FIG. 1, coil 106, receiver 108; FIG. 2, data reception unit 231; FIGS. 5A-5D; [0011], [0035], “reception coil includes a plurality of channels …”; [0037], “… a position and a patter in a k-space …”; [0049]; [0056]; [0058], “k-space data is equidistantly undersampled …”; [0061]);(b) reconstructing coil channel images from the undersampled k-space data using the computer system, wherein each coil channel image corresponds to a different channel of the multichannel receiver and each coil channel image contains aliasing artifacts (FIGS. 1-4, 9-14; [0006], “artifacts and noise remain”; [0014], “aliasing”; [0010]; [0035], “each channel … each of the reception coils”; [0037], “The measurement data for each channel corresponding to the small-sized reception coils is arranged in a k-space …”; [0089], “creates a reconstructed image … for each channel …”; [0064], “aliasing and the noise are reduced”; [0082]); (d) generating denoised k-space data from the denoised coil channel images using the computer system to transform the denoised coil channel images into k-space (FIG. 2, unit 238; FIG. 3, S304, S3042; FIGS. 4, 9-12, 14; [0013], “… creating denoised k-space data for each of the reception coils by using a denoised image …”); and (e) reconstructing denoised magnetic resonance images from the denoised k-space data using the computer system (FIGS, 2-3; FIG. 4, reconstructed image; FIGS. 9-14).
Amemiya does not disclose (c) generating denoised coil channel images with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the coil channel images using the computer system;
In the same field of endeavor, Moeller teaches a method for controlling noise in magnetic resonance imaging (Moeller: Abstract; FIGS. 1-5). Moeller further teaches generating denoised coil channel images with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the coil channel images using the computer system (Moeller: FIGS. 3A, 4; [0020], “A local low rank denoising technique may then be applied … using the matrix and the series of images”; [0027], “enforced via singular value thresholding”; [0028]).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Moeller by using LLR in order to model correlation between images that acquired at different contrasts, and improve fidelity, especially when only a few images are available.
-Regarding claim 20, Amemiya in view of Moeller teaches the method of claim 19. The combination further teaches wherein the denoised magnetic resonance images are reconstructed from the denoised k-space data using a nonlinear reconstruction algorithm (Amemiya: FIGS. 2-4, 9-14; FIG. 7; [0023]; [0068], “a deep learning model such as a convolutional neural network (CNN) learned to remove Gaussian noise”; [0070]; FIG. 8; See also Moeller; FIG. 3A).
-Regarding claim 21, Amemiya in view of Moeller teaches the method of claim 19. The combination further teaches wherein the k-space data are undersampled k-space data (Amemiya: FIG. 1, 5A-5D; [0011]; [0044]), and the coil channel images contain aliasing artifacts (Amemiya: [0014]; [0063]).
-Regarding claim 22, Amemiya in view of Moeller teaches the method of claim 21. The combination further teaches wherein the LLR model is based on a subadditivity of matrix rank for aliased image patches in the coil channel images (Moeller: FIG. 3A; [0018]).
-Regarding claim 23, Amemiya discloses a method for reconstructing denoised magnetic resonance images, the method comprising:(a) accessing k-space data with a computer system (FIG. 1, coil 106, receiver 108; FIG. 2, data reception unit 231; [0035], “reception coil includes a plurality of channels …”; [0037], “… a position and a patter in a k-space …”; [0056]); b) generating denoised coil channel images with the computer system by applying a denoising algorithm to the coil channel images using the computer system (FIG. 2, noise removal unit 237; [0067], “denoising is performed by using a predetermined algorithm …”; FIG. 3, step S303; FIGS. 4, 7, 9-10; FIGS. 11-14) and generating denoised k-space data from the denoised coil channel images using the computer system to transform the denoised coil channel images into k-space (FIG. 2, unit 238; FIG. 3, S304, S3042; FIGS. 4, 9-12, 14; [0013], “… creating denoised k-space data for each of the reception coils by using a denoised image …”); and (c) reconstructing denoised magnetic resonance images from the denoised k-space data using the computer system (FIGS. 2-4, 9-14).
Amemiya does not disclose generating denoised k-space data with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the k-space data using the computer system. A person of ordinary skills in the art would understand that this can be achieved by first generating denoised coil channel images with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the coil channel images using the computer system, and then using above Amemiya’s method (b) to generate denoised k-space data from the denoised coil channel images using the computer system to transform the denoised coil channel images into k-space.
In the same field of endeavor, Moeller teaches a method for controlling noise in magnetic resonance imaging (Moeller: Abstract; FIGS. 1-5). Moeller further teaches generating denoised coil channel images with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the coil channel images using the computer system (Moeller: FIGS. 3A, 4; [0020], “A local low rank denoising technique may then be applied … using the matrix and the series of images”; [0027], “enforced via singular value thresholding”; [0028]).
Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to combine the teaching of Amemiya with the teaching of Moeller to generate denoised coil channel images by applying a singular value thresholding using a LLR model to the coil channel images, and then generate denoised k-space data from the denoised coil channel images using the computer system to transform the denoised coil channel images into k-space in order to model correlation between images that acquired at different contrasts, and improve fidelity, especially when only a few images are available.
-Regarding claim 24, Amemiya in view of Moeller teaches the method of claim 23. The combination further teaches wherein the denoised magnetic resonance images are reconstructed from the denoised k-space data using a nonlinear reconstruction algorithm (Amemiya: FIGS. 2-4, 9-14; FIG. 7; [0023]; [0068], “a deep learning model such as a convolutional neural network (CNN) learned to remove Gaussian noise”; [0070]; FIG. 8; See also Moeller; FIG. 3A).
-Regarding claim 25, Amemiya in view of Moeller teaches the method of claim 24. The combination further teaches wherein reconstructing the denoised magnetic resonance images comprises: accessing a neural network with the computer system, wherein the neural network has been trained on training data to reconstruct magnetic resonance images from k-space data based on a nonlinear image reconstruction framework; and applying the denoised k-space data to the neural network, generating output as the denoised magnetic resonance images (Amemiya: FIGS. 2-4, 7-8; [0010], “estimated measurement data by integrating the denoised k-space data … The iterative reconstruction unit … based on the estimated measurement data”; [0070], “The output of the CNN … a denoised image …”; [0074]; [0084]; [0103], “a plurality of CNNs, such as a CNN learned with training data having a large noise amount and a CNN learned with training data having a small noise amount …”).
Response to Arguments
Applicant's arguments filed 08/05/2026 have been fully considered but they are not persuasive. Applicant argues that Amemiya does not anticipate independent claim 1 (Remarks: pages 7-8) and the combination of Amemiya and Moeller fails to teach or suggest all elements of independent claims 19 and 23 (Remarks: pages 9-11).
Regarding to claim 1 and in response to applicant’s argument “the Office's rejection requires the same CNN disclosure in Amemiya to satisfy both the denoising step (c) and the separate nonlinear-reconstruction step (e) of claim 1. This double-counting is improper because claim 1 requires two distinct operations” (Remarks: page 7 last paragraph – page 8, First paragraph), Amemiya discloses c) generating denoised coil channel images with the computer system by applying a denoising algorithm to the coil channel images using the computer system (Amemiya: FIG. 2, noise removal unit 237; [0067], "denoising is performed by using a predetermined algorithm ... "; FIG. 3, step S303; FIGS. 4, 7, 9-10; FIGS. 11-14). The predetermined algorithm includes “… a combination of wavelet transform and soft threshold processing (Approach 1), reconstruction using a dictionary representing a local structure such as sparse modeling or sparse coating (Approach 2), and a deep learning model such as a convolutional neural network (CNN) learned to remove Gaussian noise (Approach 3)” (Amemiya: [0068]). CNN can be one of the nonlinear algorithm used in this phase. Please note that Amemiya discloses denoising after synthesizing the image data on the reception coil, and the synthesis of the image data may be summing on the k-space, including convoluting a kernel corresponding to the sensitivity map with the k-space data on the respective channels and summing the channels (Amemiya: [0066]; emphasis added), thus, this phase can be considered as directly using CCN with k-space data for denoising (Amemiya’s FIG. 7, [0071], [0075]). Second, Amemiya discloses reconstructing denoised magnetic resonance images from the denoised k-space data using the computer system as shown in FIGS. 2-4, 9-14. The reconstruction is based on inverse Fourier Transform of integrated (or estimated) K-space data which is based on K-space data set with initial value and measurement data (Amemiya: FIGS. 3-4, 9-10, 12, 14). It can be anticipated that CNN and other nonlinear algorithm can be applied for reconstruction in this phase as well (Amemiya: [0110], “the processing (methods) that can be employed in each step described in these embodiments can be combined as appropriate …”). Third, Amemiya’s reconstruction method is an iterative reconstruction method by performing iterative processing while updating the initial value (i.e., denoised) based on the estimated measurement data (Amemiya: [0010]; [0013]; FIGS. 3-4, 9-10, 12, 14). Thus, the entire iterative processing is considered as reconstructing denoised magnetic resonance images from the denoised k-space data (after 1st round of iterations) using the computer system by applying the denoised k-space data to a nonlinear reconstruction algorithm associated with CNN or other nonlinear algorithms.
Regarding to claim 1 and in response to applicant’s argument “the CNN is the noise-removal step, not a reconstruction algorithm applied to denoised k-space data” (Remarks: page 8, 2nd – 3rd paragraphs), the claim does not recite details about how to apply the denoised k-space data to a nonlinear reconstruction algorithm (emphasis added). On the other hand, as explained in the above paragraph, Amemiya discloses an iterative reconstruction method by performing iterative processing while updating the initial value (i.e., denoised) based on the estimated measurement data (Amemiya: [0010]; [0013]; FIGS. 3-4, 9-10, 12, 14). Thus, the entire iterative processing is considered as reconstructing denoised magnetic resonance images from the denoised k-space data (after 1st round of iterations) using the computer system by applying the denoised k-space data to a nonlinear reconstruction algorithm associated with CNN or other nonlinear algorithms. In addition, Amemiya discloses denoising after synthesizing the image data on the iterated and updated initial values (FIGS. 3-4, 9-10, 12, 14), and the synthesis of the image data may be summing on the k-space, including convoluting a kernel corresponding to the sensitivity map with the k-space data on the respective channels and summing the channels (Amemiya: [0066]; emphasis added), thus, it can be considered as directly applying k-space data to a CNN or other nonlinear algorithm (Amemiya’s FIG. 7, [0071], [0075]).
Regarding to claim 19 and in response to applicant’s argument “The combination of Amemiya and Moeller fails to teach or suggest all elements of independent claims 19 and 23, and the Office's proposed motivation for combining Amemiya and Moeller" in order to model correlation between images that acquired at different contrasts, and improve fidelity" does not supply a rational basis for the proposed combination” (Remarks: page 9, 3rd paragraph), Moeller teaches a method for controlling noise in magnetic resonance imaging (MRI) using a locally low rank (LLR) denoise method (Moeller: Abstract; FIGS. 1-5; [0020]). The method incudes reconstructing a series of images comprises using a parallel image reconstruction (Moeller: [0026]). Parallel imaging techniques undersample k-space by reducing the number of acquired phase-encoded k-space sampling lines while keeping the maximal extent covered in k-space fixed (Moeller: [0004]). Due to redundancies in the image series that acquire the same anatomy over time or through contrast changes (e.g., across different contrast weightings), this Casorati matrix can be represented accurately by a low-rank matrix (Moeller: [0027]). Thus combination of the teaching of Amemiya and the teaching of Moeller by using LLR to model correlation between images that acquired at different contrasts will help to reduce the noise and improve the performance of reconstruction.
Regarding to claim 19 and in response to applicant’s argument “Moeller does not teach applying a singular value thresholding using an LLR model to aliased coil channel images reconstructed from undersampled k-space data, as recited” (Remarks: page 9, last paragraph), as mentioned in the above paragraph, Moeller teaches that parallel imaging techniques undersample k-space by reducing the number of acquired phase-encoded k-space sampling lines (Moeller: [0004]). Fourier representation of the signal should be sampled so that the aliasing artifacts under this transform are incoherent, and most clinically-available uniform undersampling patterns yield aliasing artifacts that are not incoherent (Moeller: [0005]). Moeller teaches further generating denoised coil channel images with the computer system by applying a singular value thresholding using a locally low-rank (LLR) model to the coil channel images using the computer system (Moeller: FIGS. 3A, 4; [0020], “A local low rank denoising technique may then be applied … using the matrix and the series of images”; [0027], “enforced via singular value thresholding”; [0028]). Please note that this is 103 rejection, Amemiya also discloses (a) accessing undersampled k-space data with a computer system, wherein the undersampled k-space data have been acquired using a multichannel receiver (FIG. 1, coil 106, receiver 108; FIG. 2, data reception unit 231; FIGS. 5A-5D; [0011], [0035], “reception coil includes a plurality of channels …”; [0037], “… a position and a patter in a k-space …”; [0049]; [0056]; [0058], “k-space data is equidistantly undersampled …”; [0061]); (b) reconstructing coil channel images from the undersampled k-space data using the computer system, wherein each coil channel image corresponds to a different channel of the multichannel receiver and each coil channel image contains aliasing artifacts (FIGS. 1-4, 9-14; [0006], “artifacts and noise remain”; [0014], “aliasing”; [0010]; [0035], “each channel … each of the reception coils”; [0037], “The measurement data for each channel corresponding to the small-sized reception coils is arranged in a k-space …”; [0089], “creates a reconstructed image … for each channel …”; [0064], “aliasing and the noise are reduced”; [0082]).
Regarding to claim 23 and in response to applicant’s arguments “claim 23 requires applying the LLR-based singular value thresholding directly to the k-space data” (Remarks: page 10 3rd paragraph), and “specification describes that "the denoised k-space data can be generated by performing denoising directly on the k-space data ... rather than on coil channel images, or other images, reconstructed from the k-space data" (Remarks: page 11, 3rd paragraph ), it is noted that the features upon which applicant relies (i.e., directly to the k-space data) are not recited in the rejected claim(s). Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). Claim 23 does not recite the details the details how to apply LLR model to the K-space data, or frequence domain data (emphasis added). It may use LLR model for the K-space data after Inverse Fourier transform, or use LLR model for the data after any other process of the K-space data, or use LLR model for absolute K-space data (real and imaginary). Please note that LLR is known method for denoising. It is a building block that can be used for any data to be denoised as shown in Moeller’s FIG. 3A.
Conclusion
THIS ACTION IS MADE FINAL. 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 XIAO LIU whose telephone number is (571)272-4539. The examiner can normally be reached Monday-Thursday and Alternate Fridays 8:30-4:30.
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, Jennifer Mehmood can be reached at (571) 272-2976. 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.
/XIAO LIU/ Primary Examiner, Art Unit 2664
Prosecution Insights