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 Objections
Claims 1 and 21 are objected to because of the following informalities:
Claim 1 lines 4-5: “using at least a part of the seismic stack images, to generate a model that is configured to reconstruct the seismic stack images” should be corrected to “using at least a part of input the seismic stack images, to generate a model that is configured to reconstruct the input seismic stack images”.
Claim 21 line 1: “wherein a cycleGAN model is used” should be corrected to “wherein a cycle generative adversarial network (cycleGAN)
Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1 and 8-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to Abstract idea without significantly more.
With respect to claim 1 the limitation(s):
accessing input seismic stack images;
performing unsupervised machine learning, using at least a part of the seismic stack images, to generate a model that is configured to reconstruct the seismic stack images;
using the model in order to generate reconstructed seismic stack images;
assessing reconstructive errors based on the reconstructed seismic stack images with the input seismic stack images;
detecting the anomalous features based on the assessment of the reconstructive errors; and
using the detected anomalous features for hydrocarbon management.
These limitation(s) highlighted in (bold) is/are directed to an abstract idea and would fall within the “Mental Processes” and “Mathematical Concepts” groupings of abstract ideas. The above portion(s) of the claim(s) constitute(s) an abstract idea because:
The limitation(s) regarding “performing unsupervised machine learning, using at least a part of the seismic stack images, to generate a model that is configured to reconstruct the seismic stack images”, as drafted, falls within the “Mathematical Concepts” groupings of abstract ideas. This interpretation is supported in the specification as shown by paragraphs [0040] – [0041] of the Specification as filed, which is an explicit recitation of an equation corresponding to the claimed limitation. It is important to note that a mathematical concept need not be expressed in mathematical symbols, because "[w]ords used in a claim operating on data to solve a problem can serve the same purpose as a formula." In re Grams, 888 F.2d 835, 837 and n.1, 12 USPQ2d 1824, 1826 and n.1 (Fed. Cir. 1989).
The limitation(s) regarding “using the model in order to generate reconstructed seismic stack images”, as drafted, is an act of observation and evaluation that, under its broadest reasonable interpretation, covers performance of the limitation(s) in the mind. That is, other than reciting “computer-implemented,” nothing in the claim language precludes the Step(s) from practically being performed in the mind. For example, but for the “computer-implemented” language, “generate” in the context of this claim encompasses the user manually generating a seismic image.
Further, the limitation regarding “using the model in order to generate reconstructed seismic stack images”, as drafted, falls within the “Mathematical Concepts” groupings of abstract ideas. This interpretation is supported in the specification as shown by paragraph [0039] of the Specification as filed, which is an explicit recitation of an equation corresponding to the claimed limitation. It is important to note that a mathematical concept need not be expressed in mathematical symbols, because "[w]ords used in a claim operating on data to solve a problem can serve the same purpose as a formula." In re Grams, 888 F.2d 835, 837 and n.1, 12 USPQ2d 1824, 1826 and n.1 (Fed. Cir. 1989).
The limitation(s) regarding “assessing reconstructive errors based on the reconstructed seismic stack images with the input seismic stack images”, as drafted, is an act of observation and evaluation that, under its broadest reasonable interpretation, covers performance of the limitation(s) in the mind. That is, other than reciting “computer-implemented,” nothing in the claim language precludes the Step(s) from practically being performed in the mind. For example, but for the “computer-implemented” language, “assessing” in the context of this claim encompasses the user manually assessing reconstructive errors.
Further, the limitation regarding “assessing reconstructive errors based on the reconstructed seismic stack images with the input seismic stack images”, as drafted, falls within the “Mathematical Concepts” groupings of abstract ideas. This interpretation is supported by the recitation of a mathematical operation acting on one or more variables to determine another. It is important to note that a mathematical concept need not be expressed in mathematical symbols, because "[w]ords used in a claim operating on data to solve a problem can serve the same purpose as a formula." In re Grams, 888 F.2d 835, 837 and n.1, 12 USPQ2d 1824, 1826 and n.1 (Fed. Cir. 1989).
The limitation(s) regarding “detecting the anomalous features based on the assessment of the reconstructive errors”, as drafted, is an act of observation and evaluation that, under its broadest reasonable interpretation, covers performance of the limitation(s) in the mind. That is, other than reciting “computer-implemented,” nothing in the claim language precludes the Step(s) from practically being performed in the mind. For example, but for the “computer-implemented” language, “detecting” in the context of this claim encompasses the user manually detecting anomalous features.
Further, referring to the MPEP 2106.04, the claim limitations are analogous to a claim to "collecting information, analyzing it, and displaying certain results of the collection and analysis," where the data analysis steps are recited at a high level of generality such that they could practically be performed in the human mind, Electric Power Group v. Alstom, S.A., 830 F.3d 1350, 1353-54, 119 USPQ2d 1739, 1741-42 (Fed. Cir. 2016).
If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Further, if a claim limitation, under its broadest reasonable interpretation, recites mathematical relationships, mathematical formulas or equations, and mathematical calculations, then it fall within the “Mathematical Concepts” groupings of abstract ideas. Accordingly, the claim recites an abstract idea.
This judicial exception is not integrated into a practical application because the non- abstract additional elements of the claims do not impose meaningful limits on practicing the abstract idea(s) recited in the preceding claim(s). In particular, the claims recited the additional elements of:
The limitation(s) regarding “accessing input seismic stack images” does/do not integrate the abstract idea into a practical application because the claim does not specify what practical application the claim is directed to. Rather the limitation is recited at such a high-level of generality that it amounts to no more than adding insignificant extra- solution activity to the judicial exception, i.e. data gathering. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they are regarded as data gathering steps necessary or routine to implement the abstract idea.
The limitation(s) regarding “using the detected anomalous features for hydrocarbon management” does/do not integrate the abstract idea into a practical application because the claim does not specify what practical application the claim is directed to. Rather the limitation is recited at such a high-level of generality that it amounts to no more than an attempt to cover any solution to an identified problem with no restriction on how the result is accomplished and no description of the mechanism for accomplishing the result, does not integrate a judicial exception into a practical application or provide significantly more because this type of recitation is equivalent to the words "apply it". Accordingly, these additional elements do not integrate the abstract idea into a practical application because they are regarded as not imposing meaningful limits on the claim such that it is not nominally or tangentially related to the invention.
The limitation(s) regarding “computer-implemented” does/do not integrate the abstract idea into a practical application because the claim limitation is a generic computer component performing the generic computer function of receiving, storing, and comparing data such that it amounts to no more than mere instruction to apply the exception using a generic computer component.
As such Examiner does NOT view that the claims:
-Improve the functioning of a computer, or to any other technology or technical field;
-Apply the judicial exception with, or by use of, a particular machine - see MPEP 2106.05(b);
-Effect a transformation or reduction of a particular article to a different state or thing - see MPEP 2106.05(c); or
-Apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception – see MPEP 2106.05(e) and Vanda Memo.
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements amount to no more than mere instructions to apply the exception using a generic computer component, or are well-understood, routine, and conventional (WURC) data gathering functions.
As discussed above with respect to integration of the abstract idea into a practical application, the additional element of “using the detected anomalous features for hydrocarbon management” is/are seen as mere instructions to apply an exception. Linking a judicial exception to a technological environment cannot provide an inventive concept. Similarly, with regards to the additional element(s) of “accessing input seismic stack images” is/are viewed as insignificant extra-solution activity, such as mere data gathering in a conventional way and, therefore, does not provide an inventive concept. Similarly, with regards to the additional element(s) of “computer-implemented” is/are view as a generic computer component performing the generic computer function of receiving, storing, and comparing data such that it amounts to no more than mere instruction to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept.
Examiner further notes that such additional elements are viewed to be well- understood, routine, and conventional (WURC) as evidenced by: Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.); Denli et al. (US 20200183047 A1); Oppert et al. (US 20110083844 A1); Salman et al. (US 20200278465 A1); Bas et al. (US 20140278115 A1); Bandura et al. (US 20190041534 A1); and Sun et al. (Sun, Qiang, John Castagna, and Zhengping Liu. "AVO inversion by artificial neural networks (ANN)." SEG International Exposition and Annual Meeting. SEG, 2000.).
Considering the claim as a whole, one of ordinary skill in the art would not know the practical application of the present invention since the claims do not apply or use the judicial exception in some meaningful way. As currently claimed, Examiner views that the additional elements do not apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, because the claims fails to recite clearly how the judicial exception is applied in a manner that does not monopolize the exception because the limitation regarding “computer-implemented,” “accessing input seismic stack images,” and “using the detected anomalous features for hydrocarbon management” can be viewed as an device, necessary data gathering, and mere instructions to apply an exception and do not impose a meaningful limitation describing what problem is being remedied or solved.
Dependent claims 8-21 when analyzed as a whole are held to be patent ineligible under 35 U.S.C. 101 because the additionally recited limitation(s) fail(s) to establish that the claim(s) is/are not directed to an abstract idea, as detailed below: there are no additional element(s) in the dependent claims that adds a meaningful limitation to the abstract idea to make the claims significantly more than the judicial exception (abstract idea).
Claims 11, 13, 14, and 16 recite limitations regarding data gathering steps and insignificant application necessary or routine to implement the abstract idea and thus are not significantly more than the abstract idea and viewed to be well known routine and conventional as evidenced by the prior art shown above.
Claims 8-12, 15-21 further limit the abstract idea with an abstract idea, such as an “Mathematical Concept” or a “Mental Process”, and thus the claims are still directed to an abstract idea without significantly more.
Claims 11 and 13 recite limitations regarding an attempt to cover any solution to an identified problem with no restriction on how the result is accomplished and no description of the mechanism for accomplishing the result, does not integrate a judicial exception into a practical application or provide significantly more because this type of recitation is equivalent to the words "apply it".
Claim 2 is seen as applying or using the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than a drafting effort designed to monopolize the exception. As such, claims 2-7 are not rejected under 35 USC 101.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-6, 8-10, 12, 14, and 16-21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.) in view of Denli et al. (US 20200183047 A1).
Regarding Claim 1. Sun teaches:
A computer-implemented method for detecting anomalous features from seismic images, the method comprising:
accessing input seismic stack images (See Fig. 6-1 and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.);
using at least a part of the seismic stack images, to generate a model that is configured to reconstruct the seismic stack images (See Fig. 3-1, Fig. 6-1 and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.);
using the model in order to generate reconstructed seismic stack images (See Fig. 6-1 and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.);
assessing reconstructive errors based on the reconstructed seismic stack images with the input seismic stack images (See Fig. 6-1 and page 50: The difference between the predicted traces and the observed far traces is an indicator of anomalous AVO behavior, possibly due to hydrocarbon reservoirs.);
detecting the anomalous features based on the assessment of the reconstructive errors (See Fig. 6-1 and page 50: The difference between the predicted traces and the observed far traces is an indicator of anomalous AVO behavior, possibly due to hydrocarbon reservoirs.); and
using the detected anomalous features for hydrocarbon management (See Fig. 6-1 and page 49: recognizing the representative anomalies for different types of hydrocarbon reservoirs.).
Sun is silent as to the language of:
performing unsupervised machine learning.
Nevertheless Denli teaches:
performing unsupervised machine learning (See para[0063]: the training of the GAN may be unconditioned or unsupervised (e.g., where the model is trained to generate realistic images from scratch.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun by performing unsupervised machine learning such as that of Denli. One of ordinary skill would have been motivated to modify Sun, because using unsupervised machine learning would have helped to generate realistic images from scratch, as recognized by Denli.
Regarding Claim 2. Sun teaches:
The method of claim 1,
wherein the unsupervised machine learning is performed so that the model is trained not to reconstruct the anomalous features competently (See page 50: Artificial neural networks are good at “learning” and “estimating” the mapping relationships between the data (Liu et al., 1998). Thus they can be trained to learn background AVO responses. Once an artificial neural network learns what is normal, it can be used to identify abnormal behavior.).
Regarding Claim 3. Sun teaches:
The method of claim 2,
wherein the anomalous features comprise anomalous features of interest and anomalous features not of interest (See Pages 53 – 54: So using the neural network inverted AVO anomaly we can discriminate the gas sand from the brine sand.); and
wherein the training to learn reconstruction is such that the model is not configured to sufficiently reconstruct the anomalous features of interest and is configured to sufficiently reconstruct the anomalous features not of interest (See Pages 53 – 54: The trained neural network learned the relationship between the near trace and far trace in the brine window. This trained neural network was used to predict far traces in the gas window. The predicted far traces were very different from the observed far traces in the gas window.).
Regarding Claim 4. Sun teaches:
The method of claim 3,
wherein the training to learn reconstruction is such that the model is not configured to sufficiently reconstruct the anomalous features of interest and is configured to sufficiently reconstruct the anomalous features not of interest comprises:
performing data preparation in order to generate additional training data associated with the anomalous features not of interest or reduce training data associated with the anomalous features of interest (See Page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.).
Regarding Claim 5. Sun teaches:
The method of claim 4,
wherein the model reconstructs the anomalous features of interest with variances thereby being unable to sufficiently reconstruct the anomalous features of interest (See Page 52 – 54: The anomalies for the training and brine windows are very small, approaching zero. The anomaly for the gas window is relatively large.);
wherein the model reconstructs the anomalous features not of interest with invariance thereby being able to sufficiently reconstruct the anomalous features not of interest (See Page 52 – 54: The anomalies for the training and brine windows are very small, approaching zero. The anomaly for the gas window is relatively large.); and
wherein the data preparation comprises sampling or data augmentation in order to generate the additional training data in order for the model to learn the invariance (See Page 93: I tried several training schemes to avoid the over training. Those schemes include increasing exit error, decreasing training time, and adding some random noise to make more learning pairs.).
Regarding Claim 6. Sun teaches:
The method of claim 5,
wherein the anomalous features not of interest comprise background (See Page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.);
wherein the data preparation comprises segmenting images into at least one zone of interest (See Page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.); and
wherein training to learn reconstruction is for the at least one zone of interest in order for the trained model to sufficiently reconstruct the background in the at least one zone of interest (See Page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.).
Regarding Claim 8. Sun teaches:
The method of claim 1,
wherein assessing the reconstructive errors comprises at least one of: (1) reconstruction loss at a pixel level (See Fig. 6-6 Fig. 6-7, and page 50: The difference between the predicted traces and the observed far traces is an indicator of anomalous AVO behavior, possibly due to hydrocarbon reservoirs.); (2) reconstruction at a latent space; or (3) generative adversarial network (GAN) rating of an anomalous feature.
Regarding Claim 9. Sun is silent as to the language of:
The method of claim 8,
wherein detecting the anomalous features based on the assessment of the reconstructive errors comprises weighting of (1), (2) and (3).
Nevertheless Denli teaches:
wherein detecting the anomalous features based on the assessment of the reconstructive errors comprises weighting of (1), (2) and (3) (See para[0084]: Various weighting of the reconstruction loss and the adversarial loss are contemplated. In particular, the weight for each of the reconstruction loss and the adversarial loss may typically range between [0,1] where 0 eliminates the impact of that loss altogether during training.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein detecting the anomalous features based on the assessment of the reconstructive errors comprises weighting of (1), (2) and (3) such as that of Denli. Denli teaches, “Thus, the weights may be selected dependent on desired quality of the generated outputs and the learning performance during training, as indicated by the loss function, for both the generator and discriminator networks” (see para[0086]). One of ordinary skill would have been motivated to modify Sun, because weighting different reconstructive errors would have helped to determine the quality of a desired output, as recognized by Denli.
Regarding Claim 10. Sun teaches:
The method of claim 1,
further comprising performing supervised machine learning (See Pages 1-2: the neural network selected for this research is a supervised feedforward multi-layer network with error backpropagation training, using a delta-learning rule.)
to generate a second model (See Page 76-77: Accordingly, the best idea is to balance the AVO anomaly before the phase unwrapping and after the phase unwrapping, use the AVO anomaly before the phase unwrapping to indicate the largest area of the hydrocarbon, and use the AVO anomaly after the phase unwrapping to indicate the most possible location of the hydrocarbon.); and
wherein detecting the anomalous features is based on both the assessment of the reconstruction errors and based on the second model (See Page 76-77: Accordingly, the best idea is to balance the AVO anomaly before the phase unwrapping and after the phase unwrapping, use the AVO anomaly before the phase unwrapping to indicate the largest area of the hydrocarbon, and use the AVO anomaly after the phase unwrapping to indicate the most possible location of the hydrocarbon.).
Regarding Claim 12. Sun teaches:
The method of claim 1,
further comprising converting the detected anomalous features into geobody objects that are characterized by a geophysical inversion method (See Page 87: Comparing with conventional inversion (sparse spike, blocky), inversion using artificial neural networks can exhibit much higher apparent resolution and can be more accurate.).
Regarding Claim 14. Sun teaches:
The method of claim 1,
wherein the input stack images are pre- or partially-stack images (See Page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.).
Regarding Claim 16. Sun teaches:
The method of claim 14,
wherein one or more pre-stack input images are used to construct other pre-stack images (See Fig. 2-3, Fig. 3-1, and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.); or
wherein all pre-stack images are inputs to the model and all pre-stack images are outputs from the model (See Pages 13-14: When all the weight in the network been adjusted, the network will calculate the output again with the new weights. This output will be compared with the desired output, if the error is smaller 13 than a given level, the training will be halted, otherwise a new iteration of adjusting weights will begin.).
Regarding Claim 17. Sun is silent as to the language of:
The method of claim 1,
wherein the unsupervised machine learning is constrained to a geologic context where anomalous features are defined.
Nevertheless Denli teaches:
wherein the unsupervised machine learning is constrained to a geologic context where anomalous features are defined (See para[0040]: Conditioning data refers a collection of data or dataset to constraint, infer or determine one or more reservoir or stratigraphic models. Conditioning data might include geophysical models, petrophysical models, seismic images.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein the unsupervised machine learning is constrained to a geologic context where anomalous features are defined such as that of Denli. Denli teaches, “The constructed reservoir models may be later conditioned 170 to comply with seismic data by adjusting their geological parameters or reservoir stratigraphic configurations” (see para[0011]). One of ordinary skill would have been motivated to modify Sun, because constraining the machine learning to a geologic context would have helped to adjust the model to comply with seismic data, as recognized by Denli.
Regarding Claim 18. Sun teaches:
The method of claim 17,
wherein the geologic context is based on at least one of geologic age, zone, environment of deposition, depth or facies (See page 68: In order to decrease the influences of depth and geological structure, the seismic data were flattened along the reservoir top.).
Regarding Claim 19. Sun teaches:
The method of claim 1,
wherein inputs to the model include geophysical inversion results (See page 104: the output error is propagated back ward.), depth (See page 5: Thus, all the seismic data points within a time window equal to the length of the wavelet are used to predict a single rock property at a given depth.), geologic zone (See page 4: An ANN was trained to learn the relationship between near-offset and the far-offset traces in a zone that doesn’t contain hydrocarbons.), environment of deposition (See page 4: An ANN was trained to learn the relationship between near-offset and the far-offset traces in a zone that doesn’t contain hydrocarbons.), and the input seismic stack images (See Fig. 2-3, Fig. 3-1, and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.).
Sun is silent as to the language of:
geologic age.
Nevertheless Denli teaches:
geologic age (See page[0012]: the horizons are unfaulted and unfolded to an isochronal geologic state which corresponds to the geologic horizon of the same age.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun using geologic age such as that of Denli. Denli teaches, “The constructed reservoir models may be later conditioned 170 to comply with seismic data by adjusting their geological parameters or reservoir stratigraphic configurations” (see para[0011]). One of ordinary skill would have been motivated to modify Sun, because constraining the machine learning to a geologic age would have helped to adjust the model to comply with seismic data, as recognized by Denli.
Regarding Claim 20. Sun is silent as to the language of:
The method of claim 1,
wherein the model is based on autoencoders, autoencoders with skip layers, generative adversarial networks, recurrent networks, transformer networks, or normalizing flow networks.
Nevertheless Denli teaches:
wherein the model is based on autoencoders (See para[0078]: autoencoder), autoencoders with skip layers, generative adversarial networks (See para[0078]: VAE-combined GAN(VAEGAN) model.), recurrent networks, transformer networks, or normalizing flow networks.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein the model is based on autoencoders, autoencoders with skip layers, generative adversarial networks, recurrent networks, transformer networks, or normalizing flow networks such as that of Denli. Denli teaches, “a machine learning process such as a decoder (that learns a mapping from a latent space to the image space)” (see para[0078]). One of ordinary skill would have been motivated to modify Sun, because using an autoencoder would have helped to map seismic data from a latent space to an image space and back, as recognized by Denli.
Regarding Claim 21. Sun is silent as to the language of:
The method of claim 1,
wherein a cycleGAN model is used to learn mapping across the input seismic stack images when the input seismic stack images are unpaired.
Nevertheless Denli teaches:
wherein a cycleGAN model is used to learn mapping across the input seismic stack images when the input seismic stack images are unpaired (See para[0066]: As another example, the reservoir models input to the GAN for training a generative network need not be conditioned (e.g., cycle GANs which do not require conditioning data paired with the reservoir models).).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein a cycleGAN model is used to learn mapping across the input seismic stack images when the input seismic stack images are unpaired such as that of Denli. One of ordinary skill would have been motivated to modify Sun, because using a cycleGAN model would have helped to train a reservoir model without requiring conditioning data pair with the reservoir model, as recognized by Denli.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.) in view of Denli et al. (US 20200183047 A1) as applied to claim 5 above, and further in view of Oppert et al. (US 20110083844 A1).
Regarding Claim 7. Sun teaches:
The method of claim 5,
wherein the anomalous features of interest comprise amplitude (See Fig. 6-1 and Page 49: The normal background is defined as the amplitude variation with offset (AVO) response within a given time and space window where hydrocarbon reservoirs are known (or presumed) not to exist.);
wherein the anomalous features not of interest comprise structural anomalies (See Page 52 – 54: The trained neural network learned the relationship between the near trace and far trace in the brine window. This trained neural network was used to predict far traces in the gas window. The predicted far traces were very different from the observed far traces in the gas window.).
Sun is silent as to the language of:
wherein the data augmentation comprises rotating seismic images in order to train the model to sufficiently reconstruct the structural anomalies.
Nevertheless Oppert teaches:
wherein the data augmentation comprises rotating seismic images in order to train the model to sufficiently reconstruct the structural anomalies (See para[0009]: Preconditioning seismic data may include rotating the seismic data to quadrature phase.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein the data augmentation comprises rotating seismic images in order to train the model to sufficiently reconstruct the structural anomalies such as that of Oppert. One of ordinary skill would have been motivated to modify Sun, because rotating seismic images would have helped to precondition the seismic data, as recognize by Oppert.
Claim(s) 11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.) in view of Denli et al. (US 20200183047 A1) as applied to claim 1 above, and further in view of Salman et al. (US 20200278465 A1).
Regarding Claim 11. Sun is silent as to the language of:
The method of claim 1,
further comprising randomly sampling patches from the input seismic stack images; and
wherein training the machine learning model uses the patches from the input seismic stack images.
Nevertheless Salman teaches:
further comprising randomly sampling patches from the input seismic stack images (See para[0137]: For example, input can be one or more seismic cubes that include interpretation information where the method 1010 can process through iterating through seismic sections (e.g., inlines, crosslines, random lines) of the data to generate output.); and
wherein training the machine learning model uses the patches from the input seismic stack images (See para[0137]: For example, input can be one or more seismic cubes that include interpretation information where the method 1010 can process through iterating through seismic sections (e.g., inlines, crosslines, random lines) of the data to generate output.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun by randomly sampling patches from the input seismic stack images; and wherein training the machine learning model uses the patches from the input seismic stack images such as that of Salman. Salman teaches, “a method 1010 that can receive seismic image data with interpretations 1004 as input and that can generate training data 1008 as output” (See para[0136]). One of ordinary skill would have been motivated to modify Sun, because using randomly sampled patches would have helped to generate training data, as recognized by Salman.
Claim(s) 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.) in view of Denli et al. (US 20200183047 A1) as applied to claim 12 above, and further in view of Bas et al. (US 20140278115 A1).
Regarding Claim 13. Sun teaches:
The method of claim 12, further comprising:
using the characterized geobodies to compile user feedback regarding whether the detected anomalous features are anomalous or not; and
using the user feedback to retrain the model.
Nevertheless Bas teaches:
using the characterized geobodies to compile user feedback regarding whether the detected anomalous features are anomalous or not (See para[0104]: For each candidate, expert annotations were captured in a log file. In one implementation, the inference engine's and expert's input may jointly be used as a feedback to the system for parameter optimization and to update model parameters.); and
using the user feedback to retrain the model (See para[0104]: For each candidate, expert annotations were captured in a log file. In one implementation, the inference engine's and expert's input may jointly be used as a feedback to the system for parameter optimization and to update model parameters.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun by using the characterized geobodies to compile user feedback regarding whether the detected anomalous features are anomalous or not; and
using the user feedback to retrain the model such as that of Bas. One of ordinary skill would have been motivated to modify Sun, because using user feedback would have helped to optimize model parameters using expert input, as recognized by Bas.
Claim(s) 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sun (Sun, Qiang. Seismic inversion by artificial neural networks. The University of Oklahoma, 2001.) in view of Denli et al. (US 20200183047 A1) as applied to claim 14 above, and further in view of Bandura et al. (US 20190041534 A1).
Regarding Claim 15. Sun teaches:
The method of claim 14,
wherein the model is configured for at least one of:
the near stack image is input to the model and the far stack image is output from the model (See Fig. 6-1 and page 50: In order to teach the ANN what is normal, near offset partially stacked traces in the background window are used as the input and the far offset partially stacked traces in the same window are used as desired output in the training process.);
the near stack image and the mid stack image are input to the model and the far stack image is output from the model;
the near stack image and far stack image are input to the model and the mid stack image is output from the model; or
the near stack image, mid stack image and far stack image are input to the model and the near stack image, mid stack image and far stack image are also output from the model.
Sun is silent as to the language of:
wherein the pre- or partially-stack images comprises near stack image, mid stack image, and far stack image.
Nevertheless Bandura teaches:
wherein the pre- or partially-stack images comprises near stack image, mid stack image, and far stack image (See para[0036]: Those of skill in the art will understand that the partial stacks (near, mid, and far) are a summation of traces along the offset or angle axis for a range of offsets/angles representative of source-receiver distances that are small (near), mid-range (mid), and large (far), calculated from a pre-stack migration.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Sun wherein the pre- or partially-stack images comprises near stack image, mid stack image, and far stack image such as that of Bandura. One of ordinary skill would have been motivated to modify Sun, because using partially stack images would have helped to summarize traces along an offset angle, as recognized by Bandura.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Sun et al. (Sun, Qiang, John Castagna, and Zhengping Liu. "AVO inversion by artificial neural networks (ANN)." SEG International Exposition and Annual Meeting. SEG, 2000.) discloses determining amplitude variation with offset using difference between observed and predicted traces (See Abstract and Fig. 3).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CARTER W FERRELL whose telephone number is (571)272-0551. The examiner can normally be reached Monday - Friday 10 am - 8 pm.
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, Catherine T. Rastovski can be reached at (571) 270-0349. 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.
/CARTER W FERRELL/Examiner, Art Unit 2857
/Catherine T. Rastovski/Supervisory Primary Examiner, Art Unit 2857
Prosecution Insights