1061 results about "Tensor" patented technology

A tensor for use with a surgical navigation system is provided. The tensor comprises a first bone engaging member engageable with a first bone and a second bone engaging member engageable with a second bone. A force-applying mechanism is configured to forcibly move the first and second bone engaging members relative to one another and a sensor detects the value of the force applied by the force-applying mechanism. A transmitter communicates a parameter associated with the tensor to the surgical navigation system.

An artificial magnetic conductor is resonant at multiple resonance frequencies. The artificial magnetic conductor is characterized by an effective media model which includes a first layer and a second layer. Each layer has a layer tensor permittivity and a layer tensor permeability having non-zero elements on the main tensor diagonal only.

An artificial neural network for analyzing input data, the input data being a 3D tensor having D channels, such as D frames of a video snippet, to recognize an action therein, including: D spatial transformer modules, each generating first and second spatial transformations and corresponding first and second attention windows using only one of the D channels, and transforming first and second regions of each of the D channels corresponding to the first and second attention windows to generate first and second patch sequences; first and second CNNs, respectively processing a concatenation of the D first patch sequences and a concatenation of the D second patch sequences; and a classification network receiving a concatenation of the outputs of the first and second CNNs and the D sets of transformation parameters of the first transformation outputted by the D spatial transformer modules, to generate a predicted action class.

A cross-component superconducting gravity gradiometer sensitive to off-diagonal components of the gradient tensor includes, for each gradient axis, a pair of closely matched angular accelerometers coupled by superconducting circuitry, including sensing circuits designed to minimize the sensitivity of the instrument to angular acceleration of the platform at which the angular accelerometers are mounted; and a mode-splitting circuitry designed to reduce a nonlinear coupling of angular acceleration to the output of the gravity gradiometer and to attain the operability of the instrument in a broader range in the frequency domain.

The invention relates to the technical field of neural networks, and specifically provides a method and apparatus for quantifying and compressing a convolutional neural network. The invention aims to solve the existing problem of large loss of network performance caused by an existing method for quantifying and compressing a neural network. The method of the invention comprises the steps of obtaining a weight tensor and an input eigen tensor of an original convolutional neural network; performing fixed-point quantization on the weight tensor and the input eigen tensor based on a preset quantization bit width; and replacing the original weight tensor and the input eigen tensor with the obtained weight fixed-point representation tensor and the input feature fixed-point representation tensor to obtain a new convolutional neural network after quantization and compression of the original convolutional neural network. The method of the invention can flexibly adjust the bit width according to different task requirements and can realize quantization and compression of the convolutional neural network without adjusting the algorithm structure and the network structure so as to reduce the occupation of memory and storage resources. The invention further provides a storage apparatus and a processing apparatus, which have the above beneficial effects.

In an exemplary embodiment, a method and system is disclosed for developing a subterranean geomechanics model of a complex geological environment. The method can include estimating a pore pressure field, a stress field, a geomechanics property field, and a geological structure field from a geological concept model; geostatistically interpolating vectors and tensors from the estimated fields; and combining the results from the estimated fields and the geostatistically interpolated vectors and tensors to derive a geostatistical geomechanical model of the geological environment.

The present invention involves a quantum computing structure, comprising: one or more logical qubits, which is encoded into a plurality of superconducting qubits; and each of the logical qubits comprises at least one operating qubit and at least one ancilla qubit. Also provided is a method of quantum computing, comprising: performing encoded quantum computing operations with logical qubits that are encoded into superconducting operating qubits and superconducting ancilla qubits. The present invention further involves a method of error correction for a quantum computing structure comprising: presenting a plurality of logical qubits, each of which comprises an operating physical qubit and an ancilla physical qubit, wherein the logical states of the plurality of logical qubits are formed from a tensor product of the states of the operating and ancilla qubits; and wherein the states of the ancilla physical qubits are suppressed; and applying strong pulses to the grouping of logical qubits.

A system for tracking a moving magnetic target uses a magnetic anomaly sensing system to continually determine magnetic gradient tensors associated with the target and converts the magnetic gradient tensors to gradient contraction scalars. A processor uses the magnetic gradient tensors and gradient contraction scalars to determine a minimum of bearing and range to the target. A velocity of the target is determined using the determined bearing and range at two points in time as the target is moving. To continually track with the moving target, the processor determines adjustments in elevation and azimuth using the magnetic gradient tensors and gradient contraction scalars along with bearing, range and determined velocities of the target. The processor can include a routine that determines range while accounting for asphericity errors introduced by the aspherical nature of constant magnetic gradient contours associated with the target.

A method upsamples an image using a non-linear fully connected neural network to produce only global details of an upsampled image and interpolates the image to produce a smooth upsampled image. The method concatenates the global details and the smooth upsampled image into a tensor and applies a sequence of nonlinear convolutions to the tensor using a convolutional neural network to produce the upsampled image.

The invention discloses a method of locating a magnetic target based on a tri-axial vector magnetic sensor array. According to the method, the magnetic target is located by a magnetic gradient tensor measuring array formed by five tri-axial vector magnetic sensors at high locating precision. The specific scheme of the method comprises the following steps of: at first, measuring the magnetic gradient tensor of any point on the periphery of the magnetic target by utilizing the magnetic gradient tensor measuring array; secondly, resolving a relative distance between the magnetic target and the magnetic measuring array and relative coordinates by utilizing the magnetic tensor; and at last, locating the magnetic target according to the resolved relative distance and the relative position coordinates. The locating method has the advantages of simplicity, practicability and high locating precision.

An image processing system detects a moving person or other object of interest. The system generates a thresholded difference image by processing a video signal or other type of image signal received from a camera. The difference image is then segmented into regions bounded by vertical lines passing through the image, and silhouette candidates are identified in one or more of the regions. Tensor voting is used to determine saliency values and corresponding tangents for each of the silhouette candidates, and the resulting values and tangents are used to detect the object of interest. In an embodiment in which the object of interest is a moving person, a neck position of the moving person may be detected by analyzing a sum of x-components of tangents along a corresponding silhouette. The detected neck position may then be utilized to determine a head position and a head size for the moving person.

A system and method to determine earth formation properties by positioning a logging tool within a wellbore in the earth formation, the logging tool having a tool rotation axis and a first, a second, and a third tilted transmitter coil, and a tilted receiver coil; rotating the logging tool about the tool rotation axis; energizing each transmitter coil; measuring a coupling signal between each transmitter coil and the receiver coil for a plurality of angles of rotation; determining a coupling tensor; and determining the earth formation properties using the coupling tensor.

A magnetic anomaly sensing system uses triaxial magnetometer (TM) sensors arranged in a three-dimensional array. A processor coupled to the TM sensors generates partial gradient contraction data, and complete gradient tensor data and corresponding complete gradient contraction data. The generated data can be used to align the three-dimensional array with a magnetic target. Once the three-dimensional array is aligned with the magnetic target, the generated data can be used to uniquely determine (i) distance to the magnetic target, (ii) position of the magnetic target relative to the three-dimensional array, and (iii) the magnetic dipole moment of the magnetic target.

A computer system and method of automatically generating a Local Grid Refinement (LGR) gridded model of a gas reservoir. A geologic file includes information identifying the locations of one or more wells according to root grid cells within a volume of the earth to be modeled. User inputs specify the number of hydraulic fractures from each well, and such parameters as the fracture length, etc. User inputs also specify the number of “splits” of the root grid cells containing hydraulic fractures; those root grid cells are then split into finer resolution grid cells of increasing width within the root grid cells containing the fractures. For horizontal wells, user inputs indicate the number of splits of root grid cells containing the lateral portions of the wellbore. Non-orthogonal and complex fractures are processed by a“nested” LGR approach. Geologic properties are assigned to each grid cell, with a tensorial adjustment included for non-orthogonal fractures, and the resulting model is available for simulation.

Disclosed are methods and related devices for use with subjects with cerebral palsy or periventricular leukomalacia. In preferred embodiments, diffusion tensor imaging (DTI) is used to identify neural areas and transcranial magnetic stimulation (TMS) is used to stimulate neural pathways.

The present invention determines the adopting ratio (weight coefficient) between the high image quality processing using the tensor projection method and the high image quality processing using another method according to the degree of deviation of the input condition of the input image, and combines these processes as appropriate. This allows a satisfactory reconstruction image to be acquired even in a case of deviation from the input condition, and avoids deterioration of the high quality image due to deterioration of the reconstruction image by the projective operation.

A 5D printer, which additively manufactures an object, includes an extruder that can move linearly along three orthogonal axes and rotationally around at least one of the axes with respect to the object while depositing a material. A gantry is movable along X, Y and Z axes, and a trunnion table movable about A and B axes is mounted on the gantry. A platen is mounted on the trunnion table, and the extruder deposits the material on the platen while moving the gantry and trunnion table. A model of the object is analyzed to produce a stress tensor for the object, and the depositing is according, to the stress tensor.

A gradiometer is disclosed which has a pair of sensor bars 41, 43 supported in housings 45, 47. Transducers 71 are located adjacent the bars 41, 43 to detect movement of the bars in response to the gravity gradient tensor. At least one of the transducers 71 comprises a first coil 510 and a second coil 516 arranged in parallel and a switch 362 for proportioning current between the coils 510 and 516 so as to create a virtual coil at a position D between the coils 510 and 516.

The method for vascular elastography comprises: i) obtaining a sequence of radio-frequency (RF) images including pre-tissue-motion and post-tissue-motion images in digital form of a vessel delimited by a vascular wall; the pre-tissue-motion and post-tissue-motion images being representative of first and second time-delayed configuration, of the whole vessel; ii) partitioning both the pre-tissue-motion and post-tissue-motion images within the vascular wall into corresponding data windows; approximating a trajectory between the pre- and post-tissue-motion for corresponding data windows; and using the trajectory for each data window to compute the full strain tensor in each data window, which allow determining the Von Mises coefficient. The method can be adapted for non-invasive vascular elastography (NIVE), for non-invasive vascular micro-elastography (MicroNIVE) on small vessels, and for endovascular elastography (EVE).

Embodiments of this invention relate to a method for analysing data related to subterranean formations including collecting data from microseismic observations of a subterranean formation that is stored on a device, analysing the data using a tensor voting method, and providing an estimate of a surface of a subterranean formation. Embodiments of this invention relate to a method for analysing data related to subterranean formations including collecting data from microseismic observations of a subterranean formation, analysing the data using a tensor voting method, providing an estimate of a surface of a subterranean formation, and tailoring an aspect of an oil field service in response to the estimate.

A gravity gradiometer is disclosed which has a sensor in the form of bars (41 and 42) which are supported on a mounting (5) which has a first mount section (10) and a second mount section (20). A first flexure web (33) pivotally couples the first and second mount sections about a first axis. The second mount has a first part (25), a second part (26) and a third part (27). The parts (25 and 26) are connected by a second flexure web (37) and the parts (26 and 27) are connected by a third flexure web (35). The bars (41 and 42) are located in housings (45 and 47) and form a monolithic structure with the housings (45 and 47) respectively. The housings (45 and 47) are connected to opposite sides of the second mount section 20. The bars (41 and 42) are connected to their respective housings by flexure webs (59). Transducers (71) are located in proximity to the bars for detecting movement of the bars to in turn enable the gravitational gradient tensor to be measured. A calibration sensor is provided for sensing whether the masses are balanced at room temperature so the balance of the masses can be adjusted by adjustable screws to balance the masses for cryogenic operation of the gradiometer. The calibration sensor comprises a resonant circuit (400, 410) and an oscillator (414). The resonant circuit includes a capacitor (400) which is formed by part of the sensor mass and a space plate (405).

The present invention provides a gravity gradiometer for measuring components of the gravity gradient tensor. The gravity gradiometer comprises at least one sensor mass for movement in response to a gravity gradient and a pivotal coupling enabling the movement of the at least one sensor mass about an axis. Further, the gravity gradiometer comprises a constant charge capacitor that is arranged so that the movement of the at least one sensor mass generates a change in a voltage across the constant charge capacitor.

Aspects of the present invention are directed to computer-implemented techniques for improving the training of artificial neural networks using a reduced precision (e.g., float16) data format. Embodiments of the present invention rescale tensor values prior to performing matrix operations (such as matrix multiplication or matrix addition) to prevent overflow and underflow. To preserve accuracy throughout the performance of the matrix operations, the scale factors are defined using a novel data format to represent tensors, wherein a matrix is represented by the tuple X, where X=(a, v[.]), wherein a is a float scale factor and v[.] are scaled values stored in the float16 format. The value of any element X[i] according to this data format would be equal to a*v[i].

The invention discloses a multimodal brain network feature fusion method based on multi-task learning, and the multimodal brain network feature fusion method based on the multi-task learning includes the steps of preprocessing the obtained functional magnetic resonance imaging (fMRI) images and diffusion tensor imaging (DTI) images, registrating the preprocessed fMRI image to the standard AAL template, carrying out a fiber tracking for preprocessed DTI images, calculating fiber anisotropy (FA) value, and constructing structure connection matrix through the AAL template. Clustering coefficient of each brain area in a function connection matrix and the structure connection matrix is calculated to be regarded as function features and structure features. As two different tasks, the function features and the structure features assess an optimal feature set by solving the problem of multi-task learning optimization. The method uses information with multiple modalities complementing each other to learn simultaneously and to classify, improves the classification accuracy, solves the problems that a single task feature does not consider the correlation between features, and the fact that only one modality feature is used for pattern classification can bring to insufficient amount of information.

The present invention provides a gravity gradiometer for measuring components of the gravity gradient tensor. The gravity gradiometer comprises at least one sensor mass for movement in response to a gravity gradient and a sensor and actuator unit for generating an electrical signal in response to the movement of the at least one sensor mass and for influencing the movement of the at least one sensor mass. The gravity gradiometer also comprises an electronic circuit for simulating an impedance. The electrical circuit is arranged for amplifying the electrical signal received from the sensor and actuator unit and for directing an actuating signal to the sensor and actuator unit. The electronic circuit comprises a differential amplifiers having first and second amplifier input terminals and an amplifier output terminal and impedances Z1, Z2, Z3, at least one of the impedances have an imaginary impedance component.