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 § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 2, 7, 8, 13, and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Egorov (US 20260160911 A1) in view of Peng (US 20220221609 A1).
Regarding claim 1:
obtain first data representing waveform returns (Egorov: The seismic receivers detect and may store a time-series of samples of earth motion caused by the seismic waves. The collection of time-series of samples recorded at many receiver locations generated by a seismic source at many source locations constitutes a seismic data set. [0001])
provide, as input to one or more machine-learning models (Egorov: Machine learning models and deep learning frameworks may be also implemented to determine a velocity model from given seismic data (402). [0104]), input data frames (Egorov seismic data (402) acquired by a seismic acquisition system (100) may be arranged in a plurality of CSGs (410) to create a 3D seismic dataset. [0077]) to generate output data in terms of time, location, and velocity (Egorov: A seismic velocity model (419) is a representation of seismic velocity at a plurality of locations within a subterranean region of interest (102). [0078]; see Note 1A), wherein each input data frame includes a sampled portion of the first data (Egorov: The collection of time-series of samples recorded at many receiver locations generated by a seismic source at many source locations constitutes a seismic data set. [0001]; see Note 1B), an interpolated portion of the first data, or both; and
generate, based on the output data, one or more images (Egorov: The methods may also include forming, using the seismic processing system, a seismic image from first set of seismic signals and the second set of seismic signals, Abstract) representing structures of an observed area associated with the waveform returns (Egorov: To determine the earth structure, including the presence of hydrocarbons, the seismic data set may be processed [0002]).
Note 1A: The Examiner understands that the seismic velocity model is defined in terms of time for two reasons:
Velocity is defined as distance over time, so one of ordinary skill in the art would understand that the velocity model inherently is defined in terms of time.
Egorov teaches: “a time-domain seismic image (432) may be generated using a process called seismic migration (also referred to as simply “migration” herein) using a seismic velocity model (419)” [0079].
Note 1B: Egorov teaches in paragraph [0001] that data may be a “collection of time-series of samples”. Egorov also teaches that: “Seismic data may also be pre-processed or partially-processed data, e.g. arranged as “common shot gathers” (CSG), i.e., sorting waveforms as acquired by different receivers and having a single source location” [0073]. That is, the seismic data may be separated based on how it was received and the location. It follows that when Egorov “determine[s] a velocity model from given seismic data” as in [0104], that at least a portion of the CSG samples may be used as input data.
Egorov fails to teach:
obtain first data representing waveform returns in terms of time, location, and ray parameter;
Peng teaches:
A device comprising: one or more processors configured to:
obtain first data representing waveform returns (Peng: DUnet engine in FIG. 3 may be used to perform a deghosting task [0044]) in terms of time, location, and ray parameter (Peng: The seismic data and the mirror data are then both transformed into a Tau-P domain of the seismic data, [0044]; see Note 1C);
Note 1C: As shown in the rejection of claim 2, Peng teaches a “deghosting” task that converts seismic data to the tau-p domain. One of ordinary skill in the art would understand that the tau-p domain is defined in terms of time, location, and ray-parameter, as the definition of tau-p domain, as taught by Bates (NPL: Inversion of seismic travel times using the Tau method) is: “A powerful technique of solving the inverse problem of seismology, the Tau method, has been described recently by Bessonova et al. Limits of the function τ(p) = T(p) - pX(p), where p is the ray parameter, T the travel time and X the epicentral distance are mapped into limits in the velocity- depth plane.” (Abstract)
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Peng with Egorov. Obtaining first data representing waveform returns in terms of time, location, and ray parameter, as in Peng, would benefit the Egorov teachings by removing unnecessary data from the input: “Deghosted data is generated using a ghost delay time for each data window, so that the deghosted data has a spectrum substantially free of frequency depletions occurring in the seismic data due to receiver-side ghosts.” (Peng, [0044]).
Regarding claim 2:
Egorov in view of Peng teaches:
The device of claim 1 (as shown above), wherein to obtain the first data, the one or more processors are configured to:
obtain waveform return data representing the waveform returns (Peng: Consider now a comparison of the efficiency of seismic processing using trained Unet, Denet and DUnet engines in deghosting the same input shot illustrated in FIG. 7 [0049]) in terms of time (Peng: y-axis is arrival time [0049]), location (Peng: x-axis is offset (1-564 receivers on a variable depth streamer at 25 m interval from one another) [0049]), and amplitude (Peng: the nuances of gray represent amplitude of detected seismic wave [0049]); and
perform one or more domain transformation operations to generate the first data based on the waveform return data (Peng: The seismic data and the mirror data are then both transformed into a Tau-P domain of the seismic data [0044]).
Regarding claim 7:
Claim 7 is substantially similar to claim 1, and is therefore rejected for similar reasons. Claim 7 contains the following notable differences:
Claim 7 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng teaches the claimed device. It follows that Egorov in view of Peng teaches the corresponding method.
Regarding claim 8:
Claim 8 is substantially similar to claim 2, and is therefore rejected for similar reasons. Claim 8 contains the following notable differences:
Claim 8 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng teaches the claimed device. It follows that Egorov in view of Peng teaches the corresponding method.
Regarding claim 13:
Claim 13 is substantially similar to claim 1, and is therefore rejected for similar reasons. Claim 13 contains the following notable differences:
Claim 13 claims a non-transitory computer-readable storage device instead of a device. Egorov teaches a non-transitory computer-readable storage device: “Automated backup systems, external storage devices, or network-attached storage (NAS) can be utilized to ensure data safety.” [0040].
Regarding claim 14:
Claim 14 is substantially similar to claim 2, and is therefore rejected for similar reasons. Claim 14 contains the following notable differences:
Claim 14 claims a non-transitory computer-readable storage device instead of a device. In the rejection of claim 13, it was shown that Egorov teaches a non-transitory computer-readable storage device.
Claims 3, 4, 9, 10, 15, 16, 19, 20, and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Egorov (US 20260160911 A1) in view of Peng (US 20220221609 A1) and Yang (NPL: Deep-learning inversion: a next generation seismic velocity-model building method; from applicant’s IDS).
Regarding claim 3:
Egorov in view of Peng teaches:
The device of claim 1 (as shown above),
Egorov in view of Peng fails to teach:
wherein the one or more machine-learning models include a multi-channel neural network, and wherein each of the input data frames is provided as input to a corresponding input channel of the multi-channel neural network.
Yang teaches:
wherein the one or more machine-learning models include a multi-channel neural network (Yang: Architecture of the network used for seismic velocity inversion. Each blue and green cube corresponds to a multi-channel feature map, Pg. 38, Figure 4), and wherein each of the input data frames is provided as input to a corresponding input channel of the multi-channel neural network (Yang: To process the seismic data, we assigned different shot gathers, generated at different source locations but from the same model as channels for the input. Therefore, the number of input channels is the same as the number of sources for each model, Pg. 13, par. 2)
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Yang with Egorov in view of Peng. Inputting data frames into corresponding channels in a multi-channel neural network, as in Yang, would benefit the Egorov in view of Peng teachings by reducing data redundancy: “The multi-shot seismic data were fed into the network together to improve data redundancy” (Yang, Pg. 13, par. 2).
Regarding claim 4:
Egorov in view of Peng and Yang teaches:
The device of claim 3 (as shown above), wherein each input channel of the multi-channel neural network is associated with a corresponding ray parameter value (see Note 4A).
Note 4A: Previously in Note 1C, it was shown that Peng teaches obtaining seismic data in the tau-p domain, which defines data in terms of a ray parameter value. Furthermore, in claim 3, it was shown that Yang teaches a multi-channel neural network where each channel corresponds to seismic data input. When the teachings of Peng and Yang are combined with Egorov, it would be obvious that each input channel is associated with a corresponding ray value of the seismic data.
Similarly, the specification of the present application teaches: “the machine-learning model(s) 136 can include a multi-channel input with each channel associated with a corresponding p value of the input data frames 134,” [0048] and that “As one example, the waveform return data 120 can be received as time-domain data (e.g., in terms of time, location, and amplitude), and the preprocessor 132 can perform domain transform operation(s) (e.g., a linear Radon transformation) to represent the waveform return data 120 in a tau-p domain (e.g. in terms of tau, location, and ray parameter (p)).” [0046].
Regarding claim 9:
Claim 9 is substantially similar to claim 3, and is therefore rejected for similar reasons. Claim 9 contains the following notable differences:
Claim 9 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng teaches the claimed device. It follows that Egorov in view of Peng teaches the corresponding method.
Regarding claim 10:
Claim 10 is substantially similar to claim 4, and is therefore rejected for similar reasons. Claim 10 contains the following notable differences:
Claim 10 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng and Yang teaches the claimed device. It follows that Egorov in view of Peng and Yang teaches the corresponding method.
Regarding claim 15:
Claim 15 is substantially similar to claim 3, and is therefore rejected for similar reasons. Claim 15 contains the following notable differences:
Claim 15 claims a non-transitory computer-readable storage device instead of a device. In the rejection of claim 13, it was shown that Egorov teaches a non-transitory computer-readable storage device.
Regarding claim 16:
Claim 16 is substantially similar to claim 4, and is therefore rejected for similar reasons. Claim 16 contains the following notable differences:
Claim 16 claims a non-transitory computer-readable storage device instead of a device. In the rejection of claim 13, it was shown that Egorov teaches a non-transitory computer-readable storage device.
Regarding claim 19:
Egorov teaches:
A device comprising: one or more processors configured to:
obtain waveform return data representing waveform returns (Egorov: The seismic receivers detect and may store a time-series of samples of earth motion caused by the seismic waves. [0001]) associated with one or more shots (Seismic data may also be pre-processed or partially-processed data, e.g. arranged as “common shot gathers” (CSG) [0073]);
perform one or more domain transform operations based on the waveform return data (Egorov: Migration is a key step that transforms the processed seismic data from the time domain to the depth domain, providing a more accurate representation of subsurface structures. It helps in locating and positioning geological features accurately. [0055]);
generate, based on the one or more time-domain velocity models (Egorov: A seismic velocity model (419) is a representation of seismic velocity at a plurality of locations within a subterranean region of interest (102). [0078]; see Note 1A), one or more images representing structures of an observed area associated with the waveform returns (Egorov: Velocity analysis is crucial for accurate imaging and interpretation of subsurface structures. It involves estimating the time depth relationship of seismic reflections and determining the velocity model of the subsurface [0054]).
Note 19A: While Egorov fails to explicitly teach transforming into the tau-p domain, when combined with Peng, the frames in Egorov would be in the tau-p domain.
Egorov fails to teach:
perform one or more domain transform operations based on the waveform return data to determine tau-p domain data representing the waveform returns;
generate, based on the tau-p domain data, tau-p domain data frames, each tau-p domain data frame associated with a corresponding p value.
provide each tau-p domain data frame as input to a corresponding channel of a multi-channel convolutional neural network to generate as output one or more time-domain velocity models; and
Peng teaches:
perform one or more domain transform operations based on the waveform return data to determine tau-p domain data representing the waveform returns (Peng: The seismic data and the mirror data are then both transformed into a Tau-P domain of the seismic data, [0044]).
generate, based on the tau-p domain data, tau-p domain data frames (Peng: The seismic data and the mirror data are then both […] divided into a plurality of data windows [0044]), each tau-p domain data frame associated with a corresponding p value (see Note 19B).
Note 19B: Peng teaches a “deghosting” task that converts seismic data to the tau-p domain. One of ordinary skill in the art would understand that the each value in the tau-p domain contains a p-value, because the definition of tau-p domain, as taught by Bates (NPL: Inversion of seismic travel times using the Tau method) teaches that: “A powerful technique of solving the inverse problem of seismology, the Tau method, has been described recently by Bessonova et al. Limits of the function τ(p) = T(p) - pX(p), where p is the ray parameter, T the travel time and X the epicentral distance are mapped into limits in the velocity- depth plane.” (Abstract)
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Peng with Egorov. Obtaining first data representing waveform returns in terms of time, location, and ray parameter, as in Peng, would benefit the Egorov teachings by removing unnecessary data from the input: “Deghosted data is generated using a ghost delay time for each data window, so that the deghosted data has a spectrum substantially free of frequency depletions occurring in the seismic data due to receiver-side ghosts.” (Peng, [0044]).
Egorov in view of Peng still fails to teach:
provide each tau-p domain data frame as input to a corresponding channel of a multi-channel convolutional neural network to generate as output one or more time-domain velocity models; and
Yang teaches:
provide each tau-p domain data frame (see Note 19C) as input to a corresponding channel (Yang: To process the seismic data, we assigned different shot gathers, generated at different source locations but from the same model as channels for the input. Therefore, the number of input channels is the same as the number of sources for each model, Pg. 13, par. 2) of a multi-channel convolutional neural network (Yang: Multi-layer neural networks are computational learning architectures that propagate the input data across a sequence of linear operators and simple non-linearities. In this system, a deep convolutional neural network (CNN), proposed by LeCun et al. (2010), is implemented […]” Pg. 6, par. 2) to generate as output one or more time-domain velocity models (Yang: In this study, we proposed the use of FCN to reconstruct subsurface parameters, i.e., P-wave velocity model, directly from raw seismic data, Pg. 6, par. 3); and
Note 19C: When the teachings of Peng are combined with Egorov, it would be obvious to one of ordinary skill in the art to utilize the tau-p domain data windows with the teachings of Yang.
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Yang with Egorov in view of Peng. Inputting data frames into corresponding channels in a multi-channel neural network, as in Yang, would benefit the Egorov in view of Peng teachings by reducing data redundancy: “The multi-shot seismic data were fed into the network together to improve data redundancy” (Yang, Pg. 13, par. 2).
Regarding claim 20:
Claim 20 is substantially similar to claim 19, and is therefore rejected for similar reasons. Claim 20 contains the following notable differences:
Claim 20 claims a method instead of a device. In the rejection of claim 19, it was shown that Egorov in view of Peng and Yang teaches the claimed device. It follows that Egorov in view of Peng and Yang teaches the corresponding method.
Regarding claim 21:
Claim 21 is substantially similar to claim 19, and is therefore rejected for similar reasons. Claim 21 contains the following notable differences:
Claim 21 claims a non-transitory computer-readable storage device instead of a device. Egorov teaches a non-transitory computer-readable storage device: “Automated backup systems, external storage devices, or network-attached storage (NAS) can be utilized to ensure data safety.” [0040].
Claims 5, 11, and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Egorov (US 20260160911 A1) in view of Peng (US 20220221609 A1) and He (US 20260160916 A1).
Regarding claim 5:
Egorov in view of Peng teaches:
The device of claim 1 (as shown above),
Egorov in view of Peng fails to explicitly teach:
wherein to generate the one or more images, the one or more processors are configured to perform integration operations using the output of the one or more machine-learning models.
He teaches:
wherein to generate the one or more images, the one or more processors are configured to perform integration operations using the output of the one or more machine-learning models (He: Kirchhoff migration function may be based on an integral form of a wave equation that corresponds to pressure wave displacement and a pressure wave velocity as function of three-dimensional space and time. As such, 3D prestack Kirchhoff depth migration may be characterized as the summation of various reflection amplitudes along diffraction traveltime curves to obtain the output seismic images, [0032]; see Note 5A)
Note 5A: In claim 1, the output of the machine learning model is “in terms of time, location, and velocity”. In paragraph [0032], He teaches that output seismic images may be generated based on an integral form of a wave equation (integration operation) that corresponds to pressure wave displacement and velocity.
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of He with Egorov in view of Peng. Performing an integration operation on the output of the machine learning model, as in He, would benefit the Egorov in view of Peng teachings because of the efficiency and flexibility regarding irregular geometry of the Kirchhoff algorithm.
Regarding claim 11:
Claim 11 is substantially similar to claim 5, and is therefore rejected for similar reasons. Claim 11 contains the following notable differences:
Claim 11 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng and He teaches the claimed device. It follows that Egorov in view of Peng and He teaches the corresponding method.
Regarding claim 17:
Claim 17 is substantially similar to claim 5, and is therefore rejected for similar reasons. Claim 17 contains the following notable differences:
Claim 17 claims a non-transitory computer-readable storage device instead of a device. In the rejection of claim 13, it was shown that Egorov teaches a non-transitory computer-readable storage device.
Claims 6, 12, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Egorov (US 20260160911 A1) in view of Peng (US 20220221609 A1), Yang (NPL: Deep-learning inversion: a next generation seismic velocity-model building method; from applicant’s IDS) and Lin (NPL: RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation; from applicant’s IDS).
Regarding claim 6:
Egorov in view of Peng teaches:
The device of claim 1 (as shown above), wherein the one or more machine-learning models include:
Egorov in view of Peng fails to teach:
a multilayer contracting path where each contracting path layer includes one or more residual blocks including multiple convolution layers with one or more skip connections; and
an expanding path including at least one multipath refinement block including one or more residual blocks, one or more multi-resolution fusion blocks, and one or more chained residual pooling blocks.
Yang teaches:
a multilayer contracting path where each contracting path layer (Yang: Figure 4 shows the detailed architecture of the proposed network. It consists of a contracting path (left), Pg. 11, Architecture of the network, par. 1) includes one or more residual blocks (see Note 6B) including multiple convolution layers (Yang: Figure 1 shows a sketch of a simple FCN. In this example, migrated seismic data are used as input, which is followed by a convolutional layer, Pg. 9, par. 1) with one or more skip connections (see Note 6A); and an expanding path (Yang: and a symmetric shape of an expanding path (right) that enables precise localization, Pg. 11, Architecture of the network, par. 1).
Note 6A: Yang showcases multiple skip connections in Figure 4.
Note 6B: The specification of the present application teaches: “For example, the residual block 220 includes an activation layer 222A, and activation layer 222F, a convolution layer 224A, a convolution layer 224F (where F is an integer greater than 1).” [0053] (emphasis added), also shown in Fig. 2. The Examiner therefore interprets Yang to teach residual blocks because Yang also teaches: “In this system, a deep convolutional neural network (CNN), proposed by LeCun et al. (2010), is implemented with linear convolutions followed by non-linear activation functions.” (Pg. 6, par. 2; emphasis added).
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Yang with Egorov in view of Peng. Utilizing a multilayer contracting path, as in Yang, would benefit the Egorov in view of Peng teachings by automating extraction of key features of the seismic data: “One key characteristic of the deep-learning method is that it can automatically extract multi-layer useful features without the need for human-curated activities and initial velocity setup” (Yang, Pg. 2).
Yang fails to teach:
an expanding path including at least one multipath refinement block including one or more residual blocks, one or more multi-resolution fusion blocks, and one or more chained residual pooling blocks.
Lin teaches:
an expanding path including at least one multipath refinement block (Lin: The individual components of our multi-path refinement network architecture RefineNet, Pg. 4, Figure 3) including one or more residual blocks (Lin: Components in RefineNet employ residual connections with identity mappings, Pg. 4, Figure 3), one or more multi-resolution fusion blocks (see Note 6C), and one or more chained residual pooling blocks (see Note 6C).
Note 6C: As part of the “multi-path refinement network” in Fig. 3 on Pg. 4, Lin showcases a Multi-resolution Fusion block and a Chained Residual Pooling block.
Before the effective filing date of the claimed invention, it would have been obvious to a person having ordinary skill in the art to combine the teachings of Lin with Egorov in view of Peng and Yang. Utilizing an expanding path including at least one multipath refinement block including one or more residual blocks, one or more multi-resolution fusion blocks, and one or more chained residual pooling blocks, as in Lin, would benefit the Egorov in view of Peng and Yang teachings by preserving features at “all levels”: “Multiple stages of spatial pooling and convolution strides reduce the final image prediction typically by a factor of 32 in each dimension, thereby losing much of the finer image structure.” (Pg. 1, col. 2); “We argue that features from all levels are helpful for semantic segmentation. […] To this end we propose a novel network architecture which effectively exploits multi-level features for generating high-resolution predictions.” (Pg. 2, par. 2).
Regarding claim 12:
Claim 12 is substantially similar to claim 6, and is therefore rejected for similar reasons. Claim 12 contains the following notable differences:
Claim 12 claims a method instead of a device. In the rejection of claim 1, it was shown that Egorov in view of Peng, Yang, and Lin teaches the claimed device. It follows that Egorov in view of Peng, Yang, and Lin teaches the corresponding method.
Regarding claim 18:
Claim 18 is substantially similar to claim 6, and is therefore rejected for similar reasons. Claim 18 contains the following notable differences:
Claim 18 claims a non-transitory computer-readable storage device instead of a device. In the rejection of claim 13, it was shown that Egorov teaches a non-transitory computer-readable storage device.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Du (NPL: Deep-Learning-Based Seismic Variable-Size Velocity Model Building) uses residual blocks to “increase the depth of network to enrich the captured features”.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VINCENT ALEXANDER PROVIDENCE whose telephone number is (571)270-5765. The examiner can normally be reached Monday-Thursday 8:30-5:00.
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, King Poon can be reached at (571)270-0728. 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.
/VINCENT ALEXANDER PROVIDENCE/Examiner, Art Unit 2617 /KING Y POON/Supervisory Patent Examiner, Art Unit 2617
Prosecution Insights