Abstract: Methods and apparatuses for seismic data processing perform a least-squares reverse time migration method in which surface-attribute-independent coefficients for the surface attribute gathers are demigrated to reduce the computational cost.

Abstract: Generally discussed herein are devices, systems, and methods for wake word verification. A method can include receiving, at a server, a message from a device indicating that an utterance of a user-defined wake word was detected at the device, the message including (a) audio samples or features extracted from the audio samples and (b) data indicating the user-defined wake word, retrieving or generating, at the server, a custom decoding graph for the user-defined wake word, wherein the decoding graph and the static portion of the wake word verification model form a custom wake word verification model for the user-defined wake word, executing the wake word verification model to determine a likelihood that the wake word was uttered, and providing a message to the device indicating whether wake was uttered based on the determined likelihood.

Abstract: Generally discussed herein are devices, systems, and methods for custom wake word selection assistance. A method can include receiving, at a device, data indicating a custom wake word provided by a user, determining one or more characteristics of the custom wake word, determining that use of the custom wake word will cause more than a threshold rate of false detections based on the characteristics, rejecting the custom wake word as the wake word for accessing a personal assistant in response to determining that use of the custom wake word will cause more than a threshold rate of false detections, and setting the custom wake word as the wake word in response to determining that use of the custom wake word will not cause more than the threshold rate of false detections.

Abstract: Generally discussed herein are devices, systems, and methods for on-device detection of a wake word. A device can include a memory including model parameters that define a custom wake word detection model, the wake word detection model including a recurrent neural network transducer (RNNT) and a lookup table (LUT), the LUT indicating a hidden vector to be provided in response to a phoneme of a user-specified wake word, a microphone to capture audio, and processing circuitry to receive the audio from the microphone, determine, using the wake word detection model, whether the audio includes an utterance of the user-specified wake word, and wake up a personal assistant after determining the audio includes the utterance of the user-specified wake word.

Abstract: A method for spectral analysis of seismic data obtains imaged seismic data and generates orthogonally shifted imaged seismic data gathers. The orthogonally shifted imaged seismic data gathers are processed to generate a spectrally processed imaged seismic data. Alternatively, the imaged seismic data are obtained using a spectral processing filter that is a function of a magnitude of a total wavenumber of the imaged seismic data in three dimensions and a spatially variable velocity function.

Abstract: Multi-vintage energy mapping selects a first seismic survey data and a second seismic survey dataset from a plurality of seismic survey datasets. The first seismic survey dataset includes a set of first energies associated with a first seismic survey geometry, and the second seismic survey dataset includes a set of second energies associated with a second seismic survey geometry. The first set of energies are mapped from the first seismic survey geometry to the second seismic survey geometry, and the second set of energies are mapped from the second seismic survey geometry to the first seismic survey geometry. An updated first seismic dataset and an updated second seismic dataset are generated such that only energies from the first and second seismic datasets associated with changes in a subsurface are preserved in the updated first and second seismic datasets.

Abstract: A device, medium and method for deblending seismic data associated with a subsurface of the earth. The method includes a step of receiving seismic data S recorded with one or more seismic receivers, wherein the seismic data S includes shot recordings generated by first and second source arrays that are simultaneously actuated; a step of detecting incoherent energy of the seismic data S using a median filter; a step of replacing incoherent parts of the seismic data with a projection filter to obtain deblended data for one of the two or more source arrays; and a step of generating in a computing device an image of the subsurface based on the deblended data.

Abstract: A method and apparatus for noise attenuation. The method includes receiving seismic data associated with at least two vintages (di, dj) collected for a same subsurface, wherein the first and second vintages (di, dj) are taken at different times; calculating a set of filters (fi, fj) that minimizes an energy function (E), wherein the energy function (E) includes a term representing a 4D difference between the first and second vintages (di, dj); calculating primaries (pi, pj) corresponding to the first and second vintages (di, dj) based on the set of (fi, fj); and calculating a 4D difference (?ij) based on the primaries (pi, pj). The 4D difference (?ij) is minimized.

Abstract: Computing device, computer instructions and method for calculating an image of a subsurface based on least square migration and image de-convolution using a matching operator F. The method includes receiving seismic data d; computing a first image m of the subsurface based on the seismic data d; computing a second image h of the subsurface based on the first image m; applying a transform operation to the first and second images m and h to obtain a first transform of the first image and a second transform of the second image; calculating the matching operator F by matching the first transform of the first image to the second transform of the second image; and generating an updated image mupdated of the subsurface based on the matching operator F and the first transform of the first image.

Abstract: Multi-vintage energy mapping selects a first seismic survey data and a second seismic survey dataset from a plurality of seismic survey datasets. The first seismic survey dataset includes a set of first energies associated with a first seismic survey geometry, and the second seismic survey dataset includes a set of second energies associated with a second seismic survey geometry. The first set of energies are mapped from the first seismic survey geometry to the second seismic survey geometry, and the second set of energies are mapped from the second seismic survey geometry to the first seismic survey geometry. An updated first seismic dataset and an updated second seismic dataset are generated such that only energies from the first and second seismic datasets associated with changes in a subsurface are preserved in the updated first and second seismic datasets.

Abstract: Computing device, computer instructions and method for calculating an image of a subsurface based on least square migration and image de-convolution using a matching operator F. The method includes receiving seismic data d; computing a first image m of the subsurface based on the seismic data d; computing a second image h of the subsurface based on the first image m; applying a transform operation to the first and second images m and h to obtain a first transform of the first image and a second transform of the second image; calculating the matching operator F by matching the first transform of the first image to the second transform of the second image; and generating an updated image mupdated of the subsurface based on the matching operator F and the first transform of the first image.

Abstract: A method for spectral analysis of seismic data obtains imaged seismic data and generates orthogonally shifted imaged seismic data gathers. The orthogonally shifted imaged seismic data gathers are processed to generate a spectrally processed imaged seismic data. Alternatively, the imaged seismic data are obtained using a spectral processing filter that is a function of a magnitude of a total wavenumber of the imaged seismic data in three dimensions and a spatially variable velocity function.

Abstract: A method for processing seismic data includes receiving seismic data and a velocity model (c(x)) for a plurality of locations (x), scaling a dimension of the seismic data according to the velocity model (c(x)) to provide a velocity normalized seismic data, and generating a final image (S(x)) of the subsurface using the velocity normalized seismic data. The velocity normalized seismic data may be a reverse-time migration image (I(x,?)) corresponding to the plurality of locations (x) and a plurality of propagation distance offsets (?). The method may also include transforming the reverse-time migration image (I(x,?)) for the plurality of selected positions (x) to a wavenumber domain to provide velocity normalized wavenumber data (I(k,?)) and suppressing data components corresponding to non-physical or undefined reflection angles to provide enhanced wavenumber data (I?(k,?)) and using the enhanced wavenumber data (I?(k,?)) to generate the final image (S(x)). A corresponding apparatus is also disclosed herein.

Abstract: A non-blended dataset related to a same surveyed area as a blended dataset is used to deblend the blended dataset. The non-blended dataset may be used to calculate a model dataset emulating the blended dataset, or may be transformed in a model domain and used to derive sparseness weights, model domain masking, scaling or shaping functions used to deblend the blended dataset.

Abstract: A method and apparatus for noise attenuation. The method includes receiving seismic data associated with at least two vintages (di, dj) collected for a same subsurface, wherein the first and second vintages (di, dj) are taken at different times; calculating a set of filters (fi, fj) that minimizes an energy function (E), wherein the energy function (E) includes a term representing a 4D difference between the first and second vintages (di, dj); calculating primaries (pi, pj) corresponding to the first and second vintages (di, dj) based on the set of (fi, fj); and calculating a 4D difference (?ij) based on the primaries (pi, pj). The 4D difference (?ij) is minimized.

Abstract: Methods and devices for seismic data processing deblend seismic data gathered using simultaneous source acquisition by applying two different deblending techniques. The second deblending technique is applied to residual data obtained after applying the first deblending technique. At least one of these first and second deblending techniques uses a signal-to-noise map.

Abstract: Methods and devices for seismic data processing deblend seismic data gathered using simultaneous source acquisition by applying two different deblending techniques. The second deblending technique is applied to residual data obtained after applying the first deblending technique. At least one of these first and second deblending techniques uses a signal-to-noise map.

Abstract: A device, medium and method for deblending seismic data associated with a subsurface of the earth. The method includes a step of receiving seismic data S recorded with one or more seismic receivers, wherein the seismic data S includes shot recordings generated by first and second source arrays that are simultaneously actuated; a step of detecting incoherent energy of the seismic data S using a median filter; a step of replacing incoherent parts of the seismic data with a projection filter to obtain deblended data for one of the two or more source arrays; and a step of generating in a computing device an image of the subsurface based on the deblended data.

Abstract: Methods and devices for seismic data processing deblend seismic data gathered using simultaneous source acquisition by applying two different deblending techniques. The second deblending technique is applied to residual data obtained after applying the first deblending technique. At least one of these first and second deblending techniques uses a signal-to-noise map.

Abstract: A device, medium and method for processing seismic data associated with a subsurface of the earth. The method includes a step of receiving seismic data (d) recorded with one or more seismic receivers; a step of processing the seismic data (d) with a first processing algorithm to obtain imaged seismic data (D); a step of constructing wave-fields (W) corresponding to the imaged seismic data (D) by solving a given wave equation, wherein the wave-fields (W) have an added axis (?) which maps back spectral properties of the imaged seismic data to its pre-imaging state; and a step of processing the wave-fields (W) along the axis (?) with a second processing algorithm.