13 results about "Shift vector" patented technology

The invention provides a short-wavelength residual static correction method based on a target reconstruction algorithm, and belongs to the technical field of seismic data processing. Comprising the following steps: modeling seismic data, expressing the seismic data as a form of a plurality of waves, conforming to a target sparse model, and assuming that the target sparse model contains a plurality of targets and is accompanied by noise; target decomposition is carried out on the seismic data containing the noise and reconstruction is carried out by adopting an iteration method, one target is reconstructed each time, and residual errors after decomposition are continuously decomposed until reconstruction of all targets is completed; the minimization problem of the target reconstruction process is converted into a regularization optimization problem with constraint conditions; and solving a quadratic optimization problem about the target to obtain a minimum value depending on a shift vector, and completing short-wavelength residual static correction by solving the maximum weight from the vertex 1 to the vertex N + 1 in the K approximate graph and reconstructing the shift vector. The invention provides the residual static correction method with robustness in combination with the properties of the seismic data under the condition that the calculation cost is not increased.

A computing device performs a method of decoding video data by determining a co-located picture of the current coding unit; locating a spatial neighbor block of the current coding unit that corresponds to the co-located picture; determining a motion shift vector for the current coding unit from one or more motion vectors associated with the spatial neighbor block according to a predefined fixed order; and reconstructing a sub-block-based temporal motion vector for a respective sub-block of a plurality of sub-blocks in the current coding unit from a corresponding sub-block in the collocated picture based on the motion shift vector.

A computer-implemented method for aligning intercorrelated asynchronous time series datasets includes the steps of:(a) retrieving a first time series dataset (x) and a second time series dataset (y), the first and second time series dataset being intercorrelated,(b) segmenting each of the first and second time series dataset (x, y) into a plurality of consecutive smaller segments (xi, yi), all segments of the first and second time series dataset (x, y) having the same length,(c) determining pairs of corresponding segments by associating successive segments of the first time series dataset (x) with corresponding segments of the second time series dataset (y),(d) optimizing, for each pair of corresponding segments, a correlation function to obtain an approximation of a first times series transformation function (f1) and of a second time series transformation function (f2),(e) using the first and second time series transformation functions (f1,f2) to determine a vector of segment shifts (s) whose components contain approximations of the shifts between the first and the second segment in a pair of corresponding segments (xi, yi),(f) applying a multi-model fitting algorithm to the segment shift vector (s), said multi-model algorithm outputting a shift function (fopt) for aligning segments of each pair of corresponding segments (xi, yi), and(g) aligning the first time series dataset (x) with the second series dataset by applying said shift function (fopt) to all pairs of corresponding segments (xi, yi),wherein the first time series transformation function (f1) is parametrized by weights (w1) of a first neural network (N1) and outputs a first highly correlated time series dataset, and wherein the second time series transformation function (f2) is parametrized by weights (w2) of a second neural network (N2) and outputs a second highly correlated time series dataset.

A computing device performs a method for decoding video data comprising: determining a co-located picture of a current coding unit; locating a spatially adjacent block of the current coding unit corresponding to the co-located picture; determining a motion shift vector for the current coding unit from one or more motion vectors associated with the spatially adjacent block according to a predetermined fixed order; and reconstructing a sub-block-based temporal motion vector for a plurality of sub-blocks of the current coding unit from a corresponding sub-block within the co-located picture based on the motion shift.

The invention provides an optimization method of a communication and sensing integrated system based on hierarchical deep reinforcement learning, which comprises the following steps: constructing an intelligent reflecting surface assisted communication and sensing integrated system model which comprises a communication and sensing integrated base station, an intelligent reflecting surface, a single-antenna user and a sensing target; calculating the sensing signal-to-noise ratio of the intelligent reflector-assisted sensing integrated system model, and setting constraint conditions based on the sensing signal-to-noise ratio; constructing a target optimization problem based on a base station transmitting beam forming vector, a sensing special waveform matrix and an intelligent reflecting surface shift vector based on constraint conditions; constructing a layered deep reinforcement learning framework, and performing optimization solution on the target optimization problem by adopting the layered deep reinforcement learning framework to obtain an optimal variable; according to the optimal variable, optimizing the intelligent reflecting surface assisted sensing integrated system model; according to the invention, a layered deep reinforcement learning method is adopted to optimize the intelligent reflecting surface auxiliary sensing integrated system, the complexity of an algorithm is significantly reduced, and the system performance is improved.

A method, computing device, non-transitory computer-readable storage medium, computer program product, and method of storing a bitstream for video decoding are provided. A computing device performs a method of decoding video data by determining a collocated picture of a current coding unit, locating spatially neighboring blocks of the current coding unit corresponding to the collocated picture, determining a motion shift vector of the current coding unit from one or more motion vectors associated with the spatially neighboring blocks according to a predefined fixed order, and reconstructing a subblock-based temporal motion vector for a respective subblock of a plurality of subblocks in the current coding unit from a corresponding subblock in the collocated picture based on the motion shift vector.

This invention discloses a supercritical airfoil flow field prediction system and method based on a dual-branch neural operator, belonging to the field of airfoil flow field prediction technology. It includes a physical condition extraction module for receiving airfoil geometry and angle of attack and calculating wall distance features; a multi-order Fourier encoding module for performing multi-resolution Fourier encoding on the flow field spatial coordinates to generate coarse and fine grid Fourier features; a multilayer perceptron supernetwork module for generating scaling and shift vectors based on geometry and angle of attack; and a decoder dual-branch backbone network, including a reference multilayer perceptron network and a residual multilayer perceptron network, which respectively receive coarse and fine grid features and output low-frequency reference flow field signals and high-frequency residual flow field signals, which are then superimposed to obtain a complete flow field prediction result. This invention overcomes the spectral bias problem of traditional surrogate models through explicit frequency decoupling, achieving high accuracy in shock wave localization and surface load prediction, while significantly accelerating computation compared to CFD solvers.

Devices, systems, and methods for automatically shifting agricultural guidance lines are disclosed. A system receives a field boundary, a stored AB path with an associated first guidance width, and a target guidance width for a current operation. The system determines a shift vector comprising a direction and magnitude based at least in part on the relationship between the stored AB path and the field boundary, generates a first guidance path by applying the shift vector, generates a plurality of parallel guidance paths at spacings corresponding to the target guidance width, and commands an automatic steering system to traverse the guidance paths. The system can apply selective regional shifts based on terrain, enforce boundary clearances, reconcile prior line nudges or sensor-detected crop row positions, and synchronize datasets with cloud services to preserve consistency across operations and equipment widths.

The invention relates to the technical field of map making, and discloses a multi-element coordinated shifting method for keeping a topological relation, which comprises the following steps of: acquiring original data of a vector map, cleaning geometric coordinates, and generating preprocessed vector map data; performing topological relation classification and constraint strength assignment on the vector map data to generate the total constraint strength of each map element; meanwhile, by calculating the minimum Euclidean distance of boundary points between the map elements and combining the minimum safety distance between the elements of the map making specification, the conflict degree of any two map elements is generated; on the basis of the total constraint strength of each map element and the conflict degree of any two map elements, constructing a shift objective function, performing iterative solution by adopting a gradient descent method, and generating an optimal shift vector of the map elements so as to realize coordinated shift of the map elements; according to the method, global optimal shifting is achieved, conflict transfer is avoided, shifting efficiency is improved, and map precision and topological integrity are guaranteed.

An apparatus for decoding encoded grid data is configured to: receive one or more syntax elements in a bitstream of encoded grid data; determining an offset value based on the one or more syntax elements; determining a transformation coefficient set; applying the offset to the set of transform coefficients to determine an updated set of transform coefficients; performing inverse transformation on the updated transformation coefficient set to determine a displacement vector set; and determining a decoding grid based on the set of displacement vectors.

A computing device performs a method of decoding video data by determining a co-located picture of the current coding unit; locating a spatial neighbor block of the current coding unit that corresponds to the co-located picture; determining a motion shift vector for the current coding unit from one or more motion vectors associated with the spatial neighbor block according to a predefined fixed order; and reconstructing a sub-block-based temporal motion vector for a respective sub-block of a plurality of sub-blocks in the current coding unit from a corresponding sub-block in the collocated picture based on the motion shift vector.

A computing device performs a method of decoding video data by determining a co-located picture of the current coding unit; locating a spatial neighbor block of the current coding unit that corresponds to the co-located picture; determining a motion shift vector for the current coding unit from one or more motion vectors associated with the spatial neighbor block according to a predefined fixed order; and reconstructing a sub-block-based temporal motion vector for a respective sub-block of a plurality of sub-blocks in the current coding unit from a corresponding sub-block in the collocated picture based on the motion shift vector.

A computer-implemented method includes receiving a sequence of acoustic frames corresponding to an utterance and generating a reference speaker embedding for the utterance. The method also includes receiving a target speaker embedding for a target speaker and generating feature-wise linear modulation (FiLM) parameters including a scaling vector and a shifting vector based on the target speaker embedding. The method also includes generating an affine transformation output that scales and shifts the reference speaker embedding based on the FiLM parameters. The method also includes generating a classification output indicating whether the utterance was spoken by the target speaker based on the affine transformation output.