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 .
Response to Arguments
Applicant’s arguments, see page 2, filed 10/6/2025, with respect to specification objection have been fully considered and are persuasive. The objection of the specification has been withdrawn.
Applicant’s arguments, see page 8, filed 10/6/2025, with respect to claim objection have been fully considered and are persuasive. The objection of the claims has been withdrawn.
Applicant’s arguments, see page 8, filed 10/6/2025, with respect to 112(b) have been fully considered and are persuasive. The 112(b) of the claims has been withdrawn.
Applicant's arguments filed 10/6/2025 have been fully considered but they are not persuasive. The remark state that the previously applied reference does not perform the features of “processing the recorded k-space measurement data to generate processed k-space magnetic resonance data, the processed k-space magnetic resonance data being presented in a dedicated form; applying the provided trained reconstruction function to the processed k-space magnetic resonance data in the dedicated form, as input data, to determine output data comprising image data”. The Examiner respectfully disagrees with this assertion and will explain why briefly below.
Regarding the claimed aspect of “the processed k-space magnetic resonance data being presented in a dedicated form”, this is interpreted as k-space data being presented in a different form than the recorded k-space measurement data. The claims state that the processed k-space MR data is “presented in a dedicated form”. The dedicated form can be in a format that is not in the k-space format. For example, clam 6 states that the processing of recorded k-space measurement data includes “a Fourier transform into the image space”. In other words, Applicant’s own claims present a situation where the processed k-space data is transformed into an image space. Therefore, since the claims teach having processed k-space data into a dedicated form that can include an image space, the rejection of the claims is maintained in view of the previously applied references.
Thus, based on the above, the rejection of the claims is maintained and disclosed below.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are: processing device, first processing interface, second processing interface, first interface, second interface in claim 10.
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 102
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 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)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
(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, 4-11 and 14-17 is/are rejected under 35 U.S.C. 102(a1 and/or a2) as being anticipated by Sandino (US Pub 2020/0249300).
CLAIMS
Re claim 1: Sandino discloses a computer-implemented method for creating image data from measurement data recorded with a magnetic resonance (MR) system, the method comprising:
providing a trained reconstruction function (e.g. the deep learning ESPIRiT network is used to perform the reconstruction function after being trained, which is taught in ¶ [19].);
[0019] The present disclosure describes methods and systems for reconstructing MR images from undersampled MRI data using a deep learning-based framework that utilizes an extended coil sensitivity model to overcome model errors, such as those caused by anatomy overlap. Undersampled k-space data may be acquired, during an MRI scan, with an MM apparatus, such as the MRI apparatus shown in FIG. 1. The MRI apparatus may include one or more multi-coil receiver arrays each including a plurality of RF coils, such as the example RF coil arrays shown in FIG. 2. During the MRI scan, each receiver coil may acquire partial k-space data (due to undersampling to accelerate scan times). As shown in the example process flow of FIG. 3, the raw k-space data may be used to reconstruct multiple initial MR images and multiple coil sensitivity maps, using an ESPIRiT calibration. The multiple initial MR images and maps are then input in a deep neural network (referred to herein as DL-ESPIRiT). The DL-ESPIRiT network reconstructs multiple MR images in an iterative fashion and outputs multiple final reconstructed MR images at the end of iteration, each corresponding to a different map of the multiple sets of sensitivity maps. These final reconstructed MR images may then be combined to one MR image and displayed to a user and used for diagnosis, as shown in the example method of FIG. 6. Further details of an example DL-ESPIRiT network are shown in FIGS. 4-5. The DL-ESPIRiT network may be trained, end-to-end, by inputting artifact-free ground truth MR images and corresponding initial MR images reconstructed directly from undersampled MR data or augmented by simulated artifacts into the DL-ESPIRiT network, as shown in the example method of FIG. 7. Example MR images having various levels of different imaging artifacts, which are reconstructed via different reconstruction techniques, including the DL-ESPIRiT technique, are shown in FIGS. 8-10.
processing the recorded k-space measurement data to generate processed k-space magnetic resonance data, the processed k-space magnetic resonance data being presented in a dedicated form (e.g. an initial reconstruction is performed on the raw k-space data as a form of processing the magnetic resonance (e.g. MR) data that is presented to the neural network in a specific form. This form also identifies the number of slices of the MR data (M) as well as the dimensions, which are given as inputs into the neural network, which is taught in ¶ [37]-[44].);
[0037] Each coil element of the coil arrays is electronically coupled to the controller unit (such as controller unit 25 of FIG. 1) via a channel. In particular, each coil element can sense the MR signals and transfer the MR signal to the data acquisition unit (such as data acquisition unit 24 of FIG. 1) of the MM apparatus via the corresponding channel. The data acquisition unit then outputs digitized MR signals to the controller unit. In some examples, each individual coil element may be coupled to one channel, and each channel may only be coupled to one coil element (e.g., anterior coil array 210 may include 12 coil elements coupled to the data acquisition unit via 12 separate channels). In other examples, more than coil element may be coupled to a given channel (e.g., anterior coil array 210 may include 12 coil elements coupled to the data acquisition unit via 6 separate channels).
[0038] The MR signals acquired from the various RF coil arrays are collected in a grid of raw data, known as k-space. K-space is an array of numbers representing spatial frequencies in the MR image. In parallel imaging, the signals from multiple receiver coils (e.g., RF coil arrays), are processed simultaneously “in parallel” along separate channels. To reduce scan times in parallel imaging, the number of phase encoding steps is reduced by acquiring only partial k-space MR data (e.g., only half the lines in k-space are filled). This may be referred to herein as undersampling MRI data. Each coil exhibits a different spatial sensitivity profile, which acts as an additional spatial encoding function, and can be used to accelerate the acquisition by subsampling (e.g., undersampling) k-space and reconstructing images by using the sensitivity information. Various reconstruction techniques or algorithms, in the image domain (e.g., SENSE) or k-space domain (e.g., GRAPPA), may be implemented to estimate the missing lines of k-space and correct the aliasing overlap in parallel imaging images. These techniques may accelerate scan times by reducing the amount of data collection without aliasing. ESPIRiT combines SENSE and GRAPPA to inherit benefits from both techniques.
[0039] Referring to FIG. 3, a schematic diagram illustrating an example process flow 300 for reconstructing MM images using a deep learning (DL)-ESPIRiT network (also referred to herein as a deep learning and extended coil sensitivity network) is shown according to an exemplary embodiment. The process flow begins at 302 where a patient is put inside an MRI scanner (which may be similar to MRI apparatus 10 shown in FIG. 1) and a scan of the patient using a multi-coil receiver array of the MM scanner is performed. The DL-ESPIRiT technique discussed herein allows for flexibility in modifying the imaging model used to acquire data during the MM scan. For example, the imaging model may incorporate off-resonance information, a signal decay model, k-space symmetry with homodyne processing, and arbitrary sampling trajectories (e.g., radial, spiral, hybrid encoding, and the like). The MR signals acquired from the multi-coil receiver array are collected as raw, k-space data, as shown at 304. The k-space data at 304 may include a number of MR signals along a Kx and Ky axis, which are the spatial frequency dimensions, for a total number C of receiver coils (or groups of receiver coils) used to acquire the data. Additionally, k-space is only partially filled due to undersampling. In some embodiments, the k-space center is more densely sampled than other regions of the k-space, for the purpose of autocalibration.
[0040] The ESPIRiT calibration is performed at 306, directly on the raw k-space data in order to estimate multiple sets of coil sensitivity maps (e.g., ESPIRiT maps), as output at 308. The ESPIRiT calibration includes generating explicit coil sensitivity maps from autocalibration data collected at an autocalibration region (e.g., center of k-space). In particular, this includes assembling the raw k-space data into a matrix (known as the calibration matrix) using a sliding window throughout the autocalibration region. Each block inside the autocalibration region is a row in the calibration matrix, and columns of the calibration matrix are shifted versions of the autocalibration region. Then an ESPIRiT reconstruction operator is generated from the right singular vectors of the calibration matrix, and the sensitivity maps which are the eigenvectors of the reconstruction operator are computed via eigenvalue decomposition, each map corresponding to one set of eigenvectors. Details of the ESPIRiT calibration can be found in “ESPIRiT—An eigenvalue approach to autocalibrating parallel MM: Where SENSE meets GRAPPA,” M. Uecker et al., Magnetic Resonance in Medicine, vol. 71, no. 3, pp. 990-1001, 2014.
[0041] The number of sets of sensitivity maps is determined according to the number of eigenvectors computed from the eigenvalue decomposition. In the ideal case, there is only a single eigenvector corresponding to the absolute eigenvalue of “1” at each location and all other eigenvalues are <<1. However, errors in the acquisition may lead to multiple eigenvectors corresponding to the absolute eigenvalue of “1” or additional eigenvalues smaller than but close to “1.” The number of sensitivity maps used in reconstruction is a hyperparameter set prior to reconstruction. In some embodiments, two sets of sensitivity maps are used in the reconstruction to reduce anatomy overlaps.
[0042] The multiple sets of ESPIRiT maps are output at 308. The ESPIRiT maps are coil sensitivity maps which present a visualization of the relative weight of each coil across the spatial dimensions, X and Y, of the image. It should be understood that although 2D images are used herein as an example for illustration, the method can be applied to 3D images. Each set of coil sensitivity maps (one map for each coil) corresponds to one MR image that is reconstructed. As shown in FIG. 3, two coil sensitivity maps (e.g., M=2) are generated at 308. However, in alternate embodiments, more than two coil sensitivity maps may be generated at 308.
[0043] At 310, an initial reconstruction of MR images is performed from the raw k-space data acquired at 304 and the multiple sets of coil sensitivity maps output at 308. For example, the process at 310 may include reconstructing multiple MR images from the undersampled k-space data and the multiple sets of coil sensitivity maps, where each initial reconstructed MR image output at 312 corresponds to a different set of the multiple sets of coil sensitivity maps. These initial MR images may be zero-filled images reconstructed based on the undersampled k-space data alone, without filling in the missing lines of k-space. Thus, the initial MR images reconstructed at 310 and output at 312 are relatively fast to compute and may be heavily aliased.
[0044] The initial MR images (two shown in the example of FIG. 3) output at 312 are then input, along with the multiple sets of coil sensitivity maps output at 308, into the DL-ESPIRiT network at 314, which may also be referred to herein as the deep learning and extended coil sensitivity network or framework. Details on the DL-ESPIRiT network are shown in FIGS. 4 and 5, as discussed further below. Generally, the DL-ESPIRiT network at 314 includes a deep neural network (which may be a convolutional neural network, in one embodiment) interspersed with data consistency layers which utilize the multiple sets of coil sensitivity maps.
applying the provided trained reconstruction function to the processed k-space magnetic resonance data in the dedicated form, as input data, to determine output data comprising image data (e.g. the DL-ESPIRiT network is trained in order to provide a reconstruction function on the input data that is processed k-space data into specific form. The initial reconstructed input data is used to determine the reconstructed MR image, which is taught in ¶ [46]-[48].); and
[0046] In this disclosure, the prior on the set of images x is modeled with a convolutional neural network (CNN), as shown in FIG. 3, which replaces the proximal operation S.sub.R in Equation 2. This gives the following equation for the DL-ESPIRiT network:
x.sup.(k+1)=CNN.sup.(k)(x.sup.(k)−A.sup.H(Ax.sup.(k)−y)), (Equation 3)
The prior information is then implicitly learned by unrolling Equation 3 and trained end-to-end as a deep CNN. Network weights are allowed to vary between unrolled iterations to enhance the network's representational power. The network is summarized in FIG. 4 and expanded on in more detail in FIG. 5. As shown in FIG. 3, the final reconstructed MR images (one corresponding to each set of coil sensitivity maps) at the end of the iteration are output from the DL-ESPIRiT network at 316. These final reconstructed MR images can then be combined to be one image which has reduced artifacts (e.g., anatomy overlap, motion, chemical shift, distortion, gradient non-linearity, and the like) compared to other reconstruction techniques, as explained further below with reference to FIGS. 8-10.
[0047] Turning to FIG. 4, a schematic diagram 400 of the DL-ESPIRiT network 314 and its inputs and outputs are shown, according to an exemplary embodiment. The inputs (initial MR images 312 and multiple sets of coil sensitivity maps (e.g., ESPIRiT maps) 308) and outputs (final reconstructed MR images 316) shown in FIG. 4 are the same as those shown in FIG. 3. As discussed above, two initial MR images 312 are input and processed simultaneously through the DL-ESPIRiT network 314 and two final reconstructed MR images 316 are output. Each of the input and output MR images corresponds to a different set of the multiple sets of coil sensitivity maps. By having multiple ESPIRiT maps and MR images, the network may split up overlapping anatomy components in the MR images, as denoted by arrows 406, and de-alias them separately. In alternate embodiments, there may be more than two sets of MR images and coil sensitivity maps (such as three, four, or the like).
[0048] The DL-ESPIRiT network 314 includes a convolutional neural network (CNN) 402 and data consistency (DC) layer 404 which are iteratively applied for a number of iterations (N). The number of iterations N can be, for example, 5, 10, 20, or any other appropriate number. The CNN and DC layer work together to reconstruct multiple MR images, each MR image corresponding to a set of coil sensitivity maps. The CNN 402 includes a plurality of convolutional layers, as discussed further below with reference to FIG. 5. The CNN 402 may also be referred to as denoising blocks. The DC layer 404 enforces consistency between input k-space data and intermediate outputs of the denoising blocks (CNN 402). This ensures that the final MR image is consistent with measured data points and consequently minimizes the chance of hallucinations. The DC layers 404 use the multiple sets of coil sensitivity maps to project back and forth between k-space and image domains. The entire DL-ESPIRiT network 314 is trained end-to-end on a loss between the output and ground truth (e.g., fully-sampled) MR images, as explained further below with reference to FIG. 7.
providing an electronic output signal representing the output data (e.g. the constructed output of the network can be output for display to a user, which is taught in ¶ [60] and [61].).
[0060] At 610, the method includes combining the multiple MR images output from the trained deep neural network to form one reconstructed MR image.
[0061] Combination of the final reconstructed images into the final combined reconstructed MR image could be done using a root-sum-of-squares approach:
l.sub.RSS=√{square root over (Σ.sub.k=1.sup.Nl.sub.k.sup.2)} (Equation 4)
The method may then continue to 612 to output (e.g., display) the final combined reconstructed MR image to a user. In one example, outputting the final combined reconstructed MR image includes displaying the final combined reconstructed MR image to a user via a display screen of a display device. In one example, the display device is display unit 33 of MM apparatus 10 shown in FIG. 1. In another example, outputting the final combined reconstructed MR image may additionally or alternatively include storing the final combined reconstructed MR image on a memory connected with the processor so that a user may access and process the stored image at a later time. A medical professional may then use the displayed and stored image for diagnosis.
Re claim 4: Sandino discloses the computer-implemented method as claimed in claim 1, wherein the processing of recorded k-space measurement data comprises determining a recording form in which the recorded k-space measurement data exists (e.g. the system determines that the data recorded is raw data to be converted into an initial reconstruction, which is taught in ¶ [43] and [44] above.).
Re claim 5: Sandino discloses the computer-implemented method as claimed in claim 4, wherein the processing of recorded k-space measurement data comprises processing steps selected based on the recording form, in which the recorded k-space measurement data exists, the dedicated form of the processed k-space magnetic resonance data being selected such that the processing steps transfer the recorded k-space measurement data from its recording form into the dedicated form of the processed k-space magnetic resonance data (e.g. the invention discloses acquiring raw k-space data from the MR scan. The raw data is also used to acquire coil sensitivity maps that correspond to a MR image to be reconstructed. Once the initial reconstruction occurs, the data is sent to the DL-ESPIRiT for processing by the neural network, which is taught in ¶ [43], [44] and [46]-[48] above.).
Re claim 6. Sandino discloses the computer-implemented method as claimed in claim 1, wherein the processing of recorded k-space measurement data comprises:
applying a regridding method,
a slice selection method,
a Fourier transform into the image space,
a Fourier transform of data present in the image space into the k-space,
a parallel acquisition techniques (PAT) method (e.g. SENSE and GRAPPA are used in the processing of MR data, which is taught in ¶ [17] and [18] above.), and/or
a correction method.
[0017] Magnetic resonance imaging (MM) is a flexible diagnostic tool that enables non-invasive visualization of soft-tissue anatomy and physiology. However, the MRI acquisition process is inherently slow, limiting its clinical application in certain cases. Scan times during an MRI scan may be reduced by undersampling, or collecting less k-space data. However, undersampling may result in aliasing artifacts that may obscure relevant anatomy. Advanced MR image reconstruction techniques such as parallel imaging can dramatically accelerate scan times by reducing the amount of data collection needed to reconstruct MR images without aliasing. SENSE (sensitivity encoding) utilizes explicit knowledge of coil array sensitivities to spatially localize signals and de-alias undersampled images. GRAPPA (generalized autocalibrating partial parallel acquisition) exploits local correlations across coils in k-space to synthesize missing data samples.
[0018] Each of these approaches have tradeoffs and another approach, termed ESPIRiT, combines SENSE and GRAPPA to inherit benefits from both techniques. ESPIRiT uses a flexible coil sensitivity model, which can incorporate non-Cartesian sampling trajectories and arbitrary image priors. ESPIRiT is robust to artifacts that arise from inconsistent coil sensitivity maps by using an extended coil sensitivity model which employs multiple sets of coil sensitivity maps. For example, objects that are larger than the prescribed field of view (FOV) can overlap and create discontinuities in sensitivity maps resulting in ghosting along the phase encoding direction. However, ESPIRiT is able to represent overlapping anatomies with multiple sets of coil sensitivity maps (as compared to only a single set of coil sensitivity maps), allowing overlapping components to be de-aliased separately from each other. Details of the ESPIRiT approach were described in “ESPIRiT—An eigenvalue approach to autocalibrating parallel MM: Where SENSE meets GRAPPA,” M. Uecker et al., Magnetic Resonance in Medicine, vol. 71, no. 3, pp. 990-1001, 2014.
Re claim 7: Sandino discloses the computer-implemented method as claimed in claim 1, wherein the trained reconstruction function is a variational neural network, an unrolled neural network, and/or a U-shaped neural network; and/or the trained reconstruction function includes a U net (e.g. a U-net can be used for the neural network, which is taught in ¶ [53].).
[0053] FIG. 5 shows one embodiment of a neural network used in the DL-ESPIRiT network 314 for illustration, not for limitation. In alternate embodiments, a different neural network architecture may be used for the DL-ESPIRiT network described herein. For example, different neural network structures may include residual networks (ResNets), U-Nets, autoencoder, recurrent neural networks, and fully connected networks. In yet other embodiments, the individual convolution and activation layers of the neural network may also be modified to natively support complex-valued data.
Re claim 8: Sandino discloses the computer-implemented method as claimed in claim 1, further comprising loading measured reference data, wherein:
the processing of measured k-space measurement data is based on the loaded measured reference data (e.g. the coil sensitivity maps are used to create an initial reconstruction, which is taught in ¶ [43] and [44] above.);
the loaded measured reference data has a data consistency assurance included in the trained reconstruction function (e.g. the DL-ESPIRiT contains data consistency layers that use the coil sensitivity maps, which is taught in ¶ [44] above.); and/or
the measured reference data is loaded when the trained reconstruction function receives input data in a form of undersampled magnetic resonance data, in a reconstruction of image data which is included in the trained reconstruction function (e.g. the sensitivity maps are loaded in the trained DL-ESPIRiT network after an initial construction of the MR data and the sensitivity maps. The network processing the input information in order to form a MR image, which is taught in ¶ [43], [44], [60] and [61] above.).
Re claim 9: Sandino discloses a non-transitory computer-readable storage medium with an executable program stored thereon, that when executed, instructs a processor to perform the computer-implemented method of claim 1 (e.g. a recording medium stores a program that is executed by a controller that performs the invention, which is taught in ¶ [30].).
[0030] The controller unit 25 includes a computer and a recording medium on which a program to be executed by the computer is recorded, in some embodiments. The program when executed by the computer causes various parts of the apparatus to carry out operations corresponding to pre-determined scanning. The recording medium may comprise, for example, a ROM, flexible disk, hard disk, optical disk, magneto-optical disk, CD-ROM, or non-volatile memory card. The controller unit 25 is connected to the operating console unit 32 and processes the operation signals input to the operating console unit 32 and furthermore controls the table 26, RF driver unit 22, gradient coil driver unit 23, and data acquisition unit 24 by outputting control signals to them. The controller unit 25 also controls, to obtain a desired image, the data processing unit 31 and the display unit 33 based on operation signals received from the operating console unit 32.
Re claim 10: Sandino discloses an image creation system comprising:
a processing device (interpretation: the image creation facility 15 has a processing facility 34 which can receive measurement data recorded by way of a first processing interface 31 with a magnetic resonance system, which is taught on page 20. This interpretation and its equivalents are utilized for this claim term hereinafter in the Office Action.) configured to process k-space measurement data recorded by a magnetic resonance (MR) system into processed k-space magnetic resonance data, which has a form corresponding to a dedicated form of input data of a trained reconstruction function (e.g. an initial reconstruction is performed on the raw k-space data as a form of processing the magnetic resonance (e.g. MR) data that is presented to the neural network in a specific form. This form also identifies the number of slices of the MR data given to the neural network when combined with coil sensitivity data, which is taught in ¶ [43] and [44] above.);
wherein the processing device includes:
a first processing interface (interpretation: the image creation facility 15 has a processing facility 34 which can receive measurement data recorded by way of a first processing interface 31 with a magnetic resonance system, which is taught on page 20. This interpretation and its equivalents are utilized for this claim term hereinafter in the Office Action.) configured to receive the recorded k-space measurement data (e.g. the recorded data is received at the data acquisition unit in receive mode, which is taught in ¶ [25].),
[0025] The T/R switch 20 can selectively electrically connect the RF body coil unit 15 to the data acquisition unit 24 when operating in receive mode, and to the RF driver unit 22 when operating in transmit mode. Similarly, the T/R switch 20 can selectively electrically connect one or more of the local RF coil arrays to the data acquisition unit 24 when the local RF coil arrays operate in receive mode, and to the RF driver unit 22 when operating in transmit mode. When the local RF coil arrays and the RF body coil unit 15 are both used in a single scan, for example if the local RF coil arrays are configured to receive MR signals and the RF body coil unit 15 is configured to transmit RF signals, then the T/R switch 20 may direct control signals from the RF driver unit 22 to the RF body coil unit 15 while directing received MR signals from the local RF coil arrays to the data acquisition unit 24. The RF body coil unit 15 may be configured to operate in a transmit-only mode, a receive-only mode, or a transmit-receive mode. The local RF coil arrays may be configured to operate in a transmit-receive mode or a receive-only mode.
a processor configured to process the received recorded k-space measurement data to transform the received recorded k-space measurement data into a form corresponding to the dedicated form (e.g. a data acquisition unit can be considered as a processor that processes the received measurement data into a specific type of data before sending this information to the data processing unit, which is taught in ¶ [28] and [30]. Also, the data processing unit can be considered as a processor that is used to process the data into a specific format for later reconstruction, which is taught in ¶ [32]. an initial reconstruction is performed on the raw k-space data as a form of processing the magnetic resonance (e.g. MR) data that is presented to the neural network in a specific form. This form also identifies the number of slices of the MR data (M) as well as the dimensions, which are given as inputs into the neural network, which is taught in ¶ [37]-[44] above.), and
[0028] The data acquisition unit 24 includes a preamplifier (not shown), a phase detector (not shown), and an analog/digital converter (not shown) used to acquire the MR signals received by the local RF coil arrays. In the data acquisition unit 24, the phase detector phase detects, using the output from the RF oscillator of the RF driver unit 22 as a reference signal, the MR signals received from the RF coil arrays and amplified by the preamplifier, and outputs the phase-detected analog magnetic resonance signals to the analog/digital converter for conversion into digital signals. The digital signals thus obtained are output to the data processing unit 31.
[0029] The MRI apparatus 10 includes a table 26 for placing the subject 16 thereon. The subject 16 may be moved inside and outside the imaging space 18 by moving the table 26 based on control signals from the controller unit 25. One or more of the RF coil arrays may be coupled to the table 26 and moved together with the table.
[0030] The controller unit 25 includes a computer and a recording medium on which a program to be executed by the computer is recorded, in some embodiments. The program when executed by the computer causes various parts of the apparatus to carry out operations corresponding to pre-determined scanning. The recording medium may comprise, for example, a ROM, flexible disk, hard disk, optical disk, magneto-optical disk, CD-ROM, or non-volatile memory card. The controller unit 25 is connected to the operating console unit 32 and processes the operation signals input to the operating console unit 32 and furthermore controls the table 26, RF driver unit 22, gradient coil driver unit 23, and data acquisition unit 24 by outputting control signals to them. The controller unit 25 also controls, to obtain a desired image, the data processing unit 31 and the display unit 33 based on operation signals received from the operating console unit 32.
[0032] The data processing unit 31 includes a computer and a recording medium on which a program to be executed by the computer to perform predetermined data processing is recorded. The data processing unit 31 is connected to the controller unit 25 and performs data processing based on control signals received from the controller unit 25. The data processing unit 31 is also connected to the data acquisition unit 24 and generates spectrum data by applying various image processing operations to the magnetic resonance signals output from the data acquisition unit 24.
a second processing interface (interpretation: The processed magnetic resonance data can be provided on a second processing interface 33, which is taught on page 20. This interpretation and its equivalents are utilized for this claim term hereinafter in the Office Action.) configured to provide the processed k-space magnetic resonance data as an output of the processing device (e.g. the data acquisition unit provides processed MR data as an output to the data processing unit, which is taught in ¶ [28] above.); and
a reconstruction device (interpretation: the image creation facility 15 has a reconstruction facility (reconstructor) 44 for creating image data, which can receive processed magnetic resonance data created by way of a first interface 41 with the processing facility 34 as input data, which is taught on pages 20 and 21. This interpretation and its equivalents are utilized for this claim term hereinafter in the Office Action.) configured to create image data (e.g. the data processing unit can be considered as the reconstruction device that makes the data, which is taught in ¶ [34] and [35]. The deep learning network can be an external device or cloud that contains an input interface for inputting the processed k-space, or initial reconstruction, into a DL-ESPIRIT network, which is taught in ¶ [34] and [35]. This contains an input and output of the processed data.), the reconstruction device including:
[0034] The MRI apparatus 10 may be configured with a deep neural system, or network, for reconstructing MR images from undersampled k-space data acquired via multiple receiver coils of the MM apparatus 10. For example, a trained deep neural network may be stored at the data processing unit 31. In some embodiments, the deep neural network may be implemented on an edge device (not shown) connected to the MRI apparatus 10. In some embodiments, the deep neural network may be implemented remotely, for example in a cloud in communication with the MRI apparatus 10. In some embodiments, portions of the deep neural network are implemented on different devices, such as any appropriate combination of the MRI apparatus 10, the edge device, the cloud, etc.
[0035] Different RF coil arrays may be utilized for different scanning objectives. To that end, one or more the RF coil arrays, such as RF coil array 210, may be disconnected from the MM apparatus 10, so that a different coil array may be connected to the MM apparatus 10. The RF coil arrays may be coupled to the T/R switch 20, and thus to the RF driver unit 22 and the data acquisition unit 24, via a connector and an RF port interface 21. Each RF coil array may be electrically coupled to one or more connectors (such as connector 17a-17c). The connector(s) may be plugged into the RF port interface 21 to electronically couple the RF coil array to the T/R switch 20. For example, coil array 210 may be electronically coupled to the MRI apparatus 10 by plugging connector 17c into RF port interface 21. As such, the local RF coil arrays may be easily changed.
a first interface (interpretation: the image creation facility 15 has a reconstruction facility (reconstructor) 44 for creating image data, which can receive processed magnetic resonance data created by way of a first interface 41 with the processing facility 34 as input data, which is taught on pages 20 and 21. This interpretation and its equivalents are utilized for this claim term hereinafter in the office action.) configured to receive, as input data, the processed k-space magnetic resonance data from the processing device (e.g. the data processing unit is used to receive data from the data acquisition unit that is processed, which is taught in ¶ [28] above. The DL-ESPIRIT can be at an external computer or cloud that has an interface to receive the initial reconstruction and an interface to output the reconstruction to another device for output, which is taught in ¶ [34] above.),
a reconstruction processor configured to apply the trained reconstruction function to the input data to determine output data including image data (e.g. the controller can instruct the data acquisition unit to perform reconstruction on the received processed data using a deep learning network, which is taught in ¶ [34], [35] and [46]-[48] above.), and
a second interface (interpretation: In a reconstruction unit (reconstruction processor) 42, a trained reconstruction function 100 is applied, wherein the output data comprising developing image data can be provided to a second interface 43, which is taught on page 21. This interpretation and its equivalents are utilized for this claim term hereinafter in the Office Action.) configured to provide the output data as an output of the reconstruction device (e.g. the data processing unit is able to output reconstructed image data to a display unit, which is taught in ¶ [33].).
[0033] The display unit 33 includes a display device and displays an image on the display screen of the display device based on control signals received from the controller unit 25. The display unit 33 displays, for example, an image regarding an input item about which the operator inputs operation data from the operating console unit 32. The display unit 33 also displays a slice image of the subject 16 generated by the data processing unit 31.
Re claim 11: Sandino discloses the image creation system as claimed in claim 10, wherein the processor is configured to process the recorded k-space measurement data by applying:
a regridding method, a slice selection method, a slice-GeneRalized Autocalibrating Partially Parallel Acquisition (slice-GRAPPA) method, a Fourier transform in image space, a Fourier transform of data present in the image space into k-space, a parallel acquisition techniques (PAT) method, a GeneRalized Autocalibrating Partially Parallel Acquisition (GRAPPA) or a Sensitivity Encoding (SENSE) method, a correction method, and/or a compressed sensing method (e.g. SENSE and GRAPPA are used in the processing of raw k-space MR data, which is taught in ¶ [17] and [18] above.).
Re claim 14: Sandino discloses the image creation system as claimed in claim 10, wherein the processing of recorded k-space measurement data comprises determining a recording form in which the recorded k-space measurement data exists (e.g. the system determines that the data recorded is k-space raw data to be converted into an initial reconstruction, which is taught in ¶ [43] and [44] above.).
Re claim 15: Sandino discloses the image creation system as claimed in claim 14, wherein the processing of recorded k-space measurement data comprises processing steps selected based on a recording form, in which the recorded k-space measurement data exists, the dedicated form of the processed k-space magnetic resonance data being selected such that the processing steps transfer the recorded k-space measurement data from its recording form into the dedicated form of the processed k-space magnetic resonance data (e.g. the invention discloses acquiring raw k-space data from the MR scan. The raw data is also used to acquire coil sensitivity maps that correspond to a MR image to be reconstructed. Once the initial reconstruction occurs, the data is sent to the DL-ESPIRiT for processing by the neural network, which is taught in ¶ [43], [44] and [46]-[48] above.).
Re claim 16: Sandino discloses a magnetic resonance (MR) system comprising: a scanner configured to record k-space measurement data; and the image creation system of claim 10 configured to process the recorded k-space measurement data (e.g. a scanner is used to record measurement data and process the measurement data, which is taught in ¶ [39].).
[0039] Referring to FIG. 3, a schematic diagram illustrating an example process flow 300 for reconstructing MM images using a deep learning (DL)-ESPIRiT network (also referred to herein as a deep learning and extended coil sensitivity network) is shown according to an exemplary embodiment. The process flow begins at 302 where a patient is put inside an MRI scanner (which may be similar to MRI apparatus 10 shown in FIG. 1) and a scan of the patient using a multi-coil receiver array of the MM scanner is performed. The DL-ESPIRiT technique discussed herein allows for flexibility in modifying the imaging model used to acquire data during the MM scan. For example, the imaging model may incorporate off-resonance information, a signal decay model, k-space symmetry with homodyne processing, and arbitrary sampling trajectories (e.g., radial, spiral, hybrid encoding, and the like). The MR signals acquired from the multi-coil receiver array are collected as raw, k-space data, as shown at 304. The k-space data at 304 may include a number of MR signals along a Kx and Ky axis, which are the spatial frequency dimensions, for a total number C of receiver coils (or groups of receiver coils) used to acquire the data. Additionally, k-space is only partially filled due to undersampling. In some embodiments, the k-space center is more densely sampled than other regions of the k-space, for the purpose of autocalibration.
Re claim 17: Sandino discloses a magnetic resonance (MR) system comprising:
a scanner configured to record k-space measurement data (e.g. a scanner is used to record measurement data, which is taught in ¶ [39] above.); and
a controller configured to:
receive the recorded k-space measurement data from the scanner (e.g. a data acquisition unit can be considered as a processor that processes the received raw k-space measurement data into a specific type of data before sending this information to the data processing unit, which is taught in ¶ [28] and [30] above. Also, the data processing unit can be considered as a processor that is used to process the data into a specific format for later reconstruction, which is taught in ¶ [32] above.);
process the recorded k-space measurement data to generate processed k-space magnetic resonance data, the processed k-space magnetic resonance data being present in a dedicated form (e.g. an initial reconstruction is performed on the raw k-space data as a form of processing the magnetic resonance (e.g. MR) data that is presented to the neural network in a specific form. This form also identifies the number of slices of the MR data given to the neural network, which is taught in ¶ [37]-[44] above.);
apply a trained reconstruction function to the processed k-space magnetic resonance data in the dedicated form, as input data, to determine output data comprising image data (e.g. the DL-ESPIRiT network is trained in order to provide a reconstruction function on the input data. The initial reconstructed input data is used to determine the reconstructed MR image, which is taught in ¶ [46]-[48] above.); and
provide an electronic output signal representing the output data (e.g. the constructed output of the network can be output for display to a user, which is taught in ¶ [60] and [61] above.).
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 factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 2, 3, 12 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sandino in view of Schlemper (US Pub 2020/0294287).
Re claim 2: Sandino discloses the computer-implemented method as claimed in claim 1, wherein the dedicated form of the processed k-space magnetic resonance data comprises:
specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space,
specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space,
specifications for a number of slices for which the processed k-space magnetic resonance data contains information (e.g. the number of slices of MR data is passed to the DL-ESPIRiT for further processing, which is seen in element 312 in figure 3, taught in ¶ [43] and [44] above.), and/or
specifications for admissible phase errors contained in the processed k-space magnetic resonance data.
However, Sandino fails to specifically teach the features of specifications for a sampling pattern with which the magnetic resonance data fills k- space,
specifications for an undersampling factor, with which the magnetic resonance data undersamples the k-space.
However, this is well known in the art as evidenced by Schlemper. Similar to the primary reference, Schlemper discloses MR imaging using deep learning (same field of endeavor or reasonably pertinent to the problem).
Schlemper discloses specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space (e.g. the invention uses k-space data as spatial frequency data used as an input into a neural network, which is taught in ¶ [83] and [84].),
[0083] The neural networks described herein may be configured to operate on data in any suitable domain. For example, one or more of the neural networks described herein may be configured to receive as input, data in the “sensor domain”, “spatial-frequency domain” (also known as k-space), and/or the image domain. Data in the “sensor domain” may comprise raw sensor measurements obtained by an MRI system. Sensor domain data may include measurements acquired line-by-line for a set of coordinates specified by a sampling pattern. A line of measurements may be termed a “readout” line. Each measurement may be a spatial frequency. As such, sensor domain data may include multiple readout lines. For example, if p readout lines were measured and each readout line included m samples, the sensor domain data may be organized in an m×p matrix. Knowing the k-space coordinates associated with each of the m×p samples, the sensor domain data may be re-organized into the corresponding k-space data, and may be then considered to be spatial frequency domain data. Data in the sensor domain as well as the data in k-space is spatial frequency data, but the spatial frequency data is organized differently in these two domains. Image-domain data may be obtained by applying an inverse Fourier transformation (e.g., an inverse fast Fourier transform if the samples fall on a grid) to k-space data.
[0084] In addition, it should be appreciated that the sensor domain, k-space, and image domain are not the only domains on which the neural networks described herein may operate. For example, the data in a source domain (e.g., sensor domain, k-space, or image domain) may be further transformed by an invertible transformation (e.g., 1D, 2D, or # d Fourier, Wavelet, and/or short-time Fourier transformation, etc.) to a target domain, the neural network may be configured to receive as input data in the target domain, and after completing processing, the output may be transformed back to the source domain.
specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space (e.g. the system discloses having input MR spatial frequency data using a sampling trajectory, which is taught in ¶ [83], [84] above, [75] and [76].).
[0075] In some embodiments, the input MR spatial frequency data may be under-sampled relative to a Nyquist criterion. For example, in some embodiments, the input MR spatial frequency data may include less than 90% (or less than 80%, or less than 75%, or less than 70%, or less than 65%, or less than 60%, or less than 55%, or less than 50%, or less than 40%, or less than 35%, or any percentage between 25 and 100) of the number of data samples required by the Nyquist criterion. In some embodiments, the reconstruction neural network was trained to reconstruct MR images from spatial frequency MR data under-sampled relative to a Nyquist criterion.
[0076] In some embodiments, the input MR spatial frequency data may have been obtained using a non-Cartesian (e.g., radial, spiral, rosette, variable density, Lis sajou, etc.) sampling trajectory, which may be used to accelerate MRI acquisition and/or be robust to motion by the subject.
Therefore, in view of Schlemper, it would have been obvious to one of ordinary skill at the time the invention was made to have the feature of specifications for a sampling pattern with which the magnetic resonance data fills k- space, specifications for an undersampling factor, with which the magnetic resonance data undersamples the k-space, incorporated in the device of Sandino, in order to specify different types of information as input into a neural network to aid in processing data, which improves the quality of the images obtained from the system (as stated in Schlemper ¶ [70]-[72]).
Re claim 3: However, Sandino fails to specifically teach the features of the computer-implemented method as claimed in claim 2, wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories.
However, this is well known in the art as evidenced by Schlemper. Similar to the primary reference, Schlemper discloses MR imaging using deep learning (same field of endeavor or reasonably pertinent to the problem).
Schlemper discloses wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories (e.g. the sensor domain data is considered as spatial frequency domain data that takes into account the k-space coordinates associated with the sampling patterns, which is taught in ¶ [83] and [84] above.).
Therefore, in view of Schlemper, it would have been obvious to one of ordinary skill at the time the invention was made to have the feature of wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories, incorporated in the device of Sandino, in order to specify different types of information as input into a neural network to aid in processing data, which improves the quality of the images obtained from the system (as stated in Schlemper ¶ [70]-[72]).
Re claim 12: Sandino discloses the image creation system as claimed in claim 10, wherein the dedicated form of the processed k-space magnetic resonance data comprises:
specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space,
specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space,
specifications for a number of slices for which the processed k-space magnetic resonance data contains information (e.g. the number of slices of MR data is passed to the DL-ESPIRiT for further processing, which is seen in element 312 in figure 3, taught in ¶ [43] and [44] above.), and/or
specifications for admissible phase errors contained in the processed k-space magnetic resonance data.
However, Sandino fails to specifically teach the features of specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space,
specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space.
However, this is well known in the art as evidenced by Schlemper. Similar to the primary reference, Schlemper discloses MR imaging using deep learning (same field of endeavor or reasonably pertinent to the problem).
Schlemper discloses specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space (e.g. the invention uses k-space data as spatial frequency data used as an input into a neural network, which is taught in ¶ [83] and [84] above.),
specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space (e.g. the system discloses having input MR spatial frequency data using a sampling trajectory, which is taught in ¶ [75], [76], [83] and [84] above.).
Therefore, in view of Schlemper, it would have been obvious to one of ordinary skill at the time the invention was made to have the feature of specifications for a sampling pattern with which the processed k-space magnetic resonance data fills k- space, specifications for an undersampling factor, with which the processed k-space magnetic resonance data undersamples the k-space, incorporated in the device of Sandino, in order to specify different types of information as input into a neural network to aid in processing data, which improves the quality of the images obtained from the system (as stated in Schlemper ¶ [70]-[72]).
Re claim 13: However, Sandino fails to specifically teach the features of the image creation system as claimed in claim 12, wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories.
However, this is well known in the art as evidenced by Schlemper. Similar to the primary reference, Schlemper discloses MR imaging using deep learning (same field of endeavor or reasonably pertinent to the problem).
Schlemper discloses wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories (e.g. the sensor domain data is considered as spatial frequency domain data that takes into account the k-space coordinates associated with the sampling patterns, which is taught in ¶ [83] and [84] above.).
Therefore, in view of Schlemper, it would have been obvious to one of ordinary skill at the time the invention was made to have the feature of wherein the specifications for the sampling pattern with which the processed k-space magnetic resonance data fills k-space comprise a dimensionality of the filled k-space and/or used k-space trajectories, incorporated in the device of Sandino, in order to specify different types of information as input into a neural network to aid in processing data, which improves the quality of the images obtained from the system (as stated in Schlemper ¶ [70]-[72]).
Claim(s) 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sandino in view of Lee (NPL titled “Deep Residual Learning for Accelerated MRI Using Magnitude and Phase Networks, Date: Sept. 2018).
Re claim 18: (New) However, Sandino fails to specifically teach the features of the image creation system as claimed in claim 10, wherein the dedicated form of the processed k-space magnetic resonance data comprises specifications for admissible phase errors contained in the processed k-space magnetic resonance data.
However, this is well known in the art as evidenced by Lee. Similar to the primary reference, Lee discloses determining a phase error (same field of endeavor or reasonably pertinent to the problem).
Lee discloses wherein the dedicated form of the processed k-space magnetic resonance data comprises specifications for admissible phase errors contained in the processed k-space magnetic resonance data (e.g. the invention discloses calculating a phase error from the MRI k-space data. The system determines which part of the phase is acceptable and what part of the phase is not, which is taught in page 4 Section C. Reconstruction Flow. After being trained, the inference performs the same process as the training.).
Therefore, in view of Lee, it would have been obvious to one of ordinary skill before the effective filing date of the claimed invention was made to have the feature of wherein the dedicated form of the processed k-space magnetic resonance data comprises specifications for admissible phase errors contained in the processed k-space magnetic resonance data, incorporated in the device of Sandino, in order to provide acceptable phase error as information used to reconstruct an image, which can result in an improve artifact removal within an image (as stated in Lee Abstract).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Zhu discloses deep learning for MRI.
Schlemper discloses a k-space conversion to another output MR spatial frequency data.
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 CHAD S DICKERSON whose telephone number is (571)270-1351. The examiner can normally be reached Monday-Friday 10AM-6PM 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, Abderrahim Merouan can be reached at 571-270-5254. 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.
/CHAD DICKERSON/ Primary Examiner, Art Unit 2682