REDES NEURONALES

VolverVolver

Resultados 228 resultados LastUpdate Última actualización 17/08/2019 [20:34:00] pdf PDF

Solicitudes publicadas en los últimos 60 días / Applications published in the last 60 days

Página1 de 10 nextPage   por página


ADAPTIVE NEURAL NETWORK SELECTION TO EXTRACT PARTICULAR RESULTS

NºPublicación: US2019251402A1 15/08/2019

Solicitante:

SLINGSHOT AEROSPACE INC [US]

Resumen de: US2019251402A1

Method, electronic device, and computer readable medium embodiments are disclosed. In one embodiment, a method includes receiving image data, manipulating the received image data based on a set of transform parameters, and analyzing the manipulated image data to generate metadata. The metadata statistically describes the received image data. The method also includes selecting a neural network from a plurality of neural networks to perform a second analysis, wherein the neural network is selected based on the generated metadata. The method additionally includes performing a second analysis of the received image data by the selected neural network based on the generated metadata to extract information from the received image data.

traducir

SOURCE SEPARATION METHOD AND SOURCE SEPERATION DEVICE

NºPublicación: US2019251421A1 15/08/2019

Solicitante:

NATIONAL CENTRAL UNIV [TW]

Resumen de: US2019251421A1

A source separation method and a source separation device are provided. The source separation method comprises: obtaining at least two source time-frequency signals and a mixed time-frequency signal of the at least two source time-frequency signals; disposing the mixed time-frequency signal at an input layer of a complex-valued deep neural network, and taking the at least two time-frequency signals as a target of the complex-valued deep neural network; calculating a cost function of the complex-valued deep neural network; and performing partial differential to a real part and an imaginary part of a network parameter of the complex-valued deep neural network respectively to minimize the cost function.

traducir

METHOD AND SYSTEM FOR ACTIVITY CLASSIFICATION

NºPublicación: US2019251340A1 15/08/2019

Solicitante:

WRNCH INC [CA]

Resumen de: US2019251340A1

This disclosure is directed to an activity classifier system, for classifying human activities using 2D skeleton data. The system includes a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data, the transformed skeleton data comprising scaled, relative joint positions and relative joint velocities. It also includes a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify the most probable of a plurality of gestures. There is also an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural networks and is trained to identify the most probable of a plurality of actions.

traducir

METHOD FOR PROCESSING INFORMATION, INFORMATION PROCESSING APPARATUS, AND NON-TRANSITORY COMPUTER-READABLE RECORDING MEDIUM

NºPublicación: US2019251383A1 15/08/2019

Solicitante:

PANASONIC IP MAN CO LTD [JP]

CN_108307660_A

Resumen de: US2019251383A1

Inputting an image to a neural network, performing convolution on a current frame included in the image to calculate a current feature map, which is a feature map at a present time, combining a past feature map, which is obtained by performing convolution on a past frame included in the image, and the current feature map, estimating an object candidate area using the combined past feature map and current feature map, estimating positional information and identification information regarding the one or more objects included in the current frame using the combined past feature map and current feature map and the estimated object candidate area, and outputting the positional information and the identification information regarding the one or more objects included in the current frame of the image estimated in the estimating as object detection results are included.

traducir

FACE DETECTION TRAINING METHOD AND APPARATUS, AND ELECTRONIC DEVICE

NºPublicación: US2019251333A1 15/08/2019

Solicitante:

TENCENT TECH SHENZHEN CO LTD [CN]

CN_108985135_A

Resumen de: US2019251333A1

An object detection training method can include receiving a training sample set in a current iteration of an object detection training process over an object detection neural network. The training sample set can include first samples of a first class and second samples of a second class. A first center loss value of each of the first and second samples can be determined. The first center loss value can be a distance between an eigenvector of the respective sample and a center eigenvector of the first or second class which the respective sample belongs to. A second center loss value of the training sample set can be determined according to the first center loss values of the first and second samples. A first target loss value of the current iteration can be determined according to the second center loss value of the training sample set.

traducir

PREDICTING PATHOLOGICAL COMPLETE RESPONSE TO NEOADJUVANT CHEMOTHERAPY FROM BASELINE BREAST DYNAMIC CONTRAST ENHANCED MAGNETIC RESONANCE IMAGING (DCE-MRI)

NºPublicación: US2019251688A1 15/08/2019

Solicitante:

UNIV CASE WESTERN RESERVE [US]

Resumen de: US2019251688A1

Embodiments access a pre-neoadjuvant chemotherapy (NAC) radiological image of a region of tissue demonstrating breast cancer (BCa), the region of tissue including a tumoral region, the image having a plurality of pixels; extract a set of patches from the tumoral region; provide the set of patches to a convolutional neural network (CNN) configured to discriminate tissue that will experience pathological complete response (pCR) post-NAC from tissue that will not; receive, from the CNN, a pixel-level localized patch probability of pCR; compute a distribution of predictions across analyzed patches based on the pixel-level localized patch probability; classify the region of tissue as a responder or non-responder based on the distribution of predictions, and display the classification. Embodiments may further generate a probability mask based on the pixel-level localized patch probability; and generate a heatmap of likelihood of response to NAC based on the probability mask and the pre-NAC radiological image.

traducir

DEEP-LEARNING-BASED AUTOMATIC SKIN RETOUCHING

NºPublicación: US2019251674A1 15/08/2019

Solicitante:

ADOBE INC [US]

Resumen de: US2019251674A1

Embodiments disclosed herein involve techniques for automatically retouching photos. A neural network is trained to generate a skin quality map from an input photo. The input photo is separated into high and low frequency layers which are separately processed. A high frequency path automatically retouches the high frequency layer using a neural network that accepts the skin quality map as an input. A low frequency path automatically retouches the low frequency layer using a color transformation generated by a second neural network and the skin quality map. The retouched high and low frequency layers are combined to generate the final output. In some embodiments, a training set for any or all of the networks is enhanced by applying a modification to an original image from a pair of retouched photos in the training set to improve the resulting performance of trained networks over different input conditions.

traducir

PRUNING CONVOLUTIONAL NEURAL NETWORKS

NºPublicación: US2019251442A1 15/08/2019

Solicitante:

NVIDIA CORP [US]

Resumen de: US2019251442A1

A neural network includes at least a first network layer that includes a first set of filters and a second network layer that includes a second set of filters. Notably, a filter was removed from the first network layer. A bias associated with a different filter included in the second set of filters compensates for a different bias associated with the filter that was removed from the first network layer.

traducir

CONDITIONAL LOSS FUNCTION MODIFICATION IN A NEURAL NETWORK

NºPublicación: US2019251398A1 15/08/2019

Solicitante:

SLINGSHOT AEROSPACE INC [US]

Resumen de: US2019251398A1

Method, electronic device, and computer readable medium embodiments are disclosed. In one embodiment, a method includes training a neural network using a first image dataset and a first truth dataset, then using the trained neural network to analyze a second image dataset. The training includes modifying a loss function of the neural network to forego penalizing the neural network when a feature is predicted with higher than a first confidence level by the neural network, and the first truth dataset has no feature corresponding to the predicted feature.

traducir

IMAGE COMPOSITES USING A GENERATIVE ADVERSARIAL NEURAL NETWORK

NºPublicación: US2019251401A1 15/08/2019

Solicitante:

ADOBE INC [US]

Resumen de: US2019251401A1

The present disclosure relates to an image composite system that employs a generative adversarial network to generate realistic composite images. For example, in one or more embodiments, the image composite system trains a geometric prediction neural network using an adversarial discrimination neural network to learn warp parameters that provide correct geometric alignment of foreground objects with respect to a background image. Once trained, the determined warp parameters provide realistic geometric corrections to foreground objects such that the warped foreground objects appear to blend into background images naturally when composited together.

traducir

METHOD FOR PROCESSING INPUT ON BASIS OF NEURAL NETWORK LEARNING AND APPARATUS THEREFOR

NºPublicación: US2019251396A1 15/08/2019

Solicitante:

SAMSUNG ELECTRONICS CO LTD [KR]

KR_20180051335_A

Resumen de: US2019251396A1

An electronic device is provided. The electronic device includes a memory storing parameter sets, each of which includes one weight and bias sets respectively corresponding to n (where, n>1) occlusion levels among a plurality of occlusion levels within a certain range and at least one processor configured to obtain output data by inputting input data to a neural network.

traducir

Determining Subsurface Layers Using Machine Learning

NºPublicación: US2019250294A1 15/08/2019

Solicitante:

SCHLUMBERGER TECHNOLOGY CORP [US]

WO_2018072815_A1

Resumen de: US2019250294A1

A method is disclosed and includes receiving a seismic cube. The seismic cube includes a three-dimensional image of a portion of a subsurface area. The method further includes providing the seismic cube to a machine learning process. The machine learning process includes one or more neural networks used for predicting a location of a subsurface seismic layer in the received seismic cube. The method also includes receiving, from the machine learning process, the prediction of the location of the subsurface seismic layer in the seismic cube.

traducir

METHOD AND ELECTRONIC DEVICE FOR GENERATING TEXT COMMENT ABOUT CONTENT

NºPublicación: US2019251355A1 15/08/2019

Solicitante:

SAMSUNG ELECTRONICS CO LTD [KR]

Resumen de: US2019251355A1

Provided are an artificial intelligence (AI) system that mimics functions, such as recognition and determination, of the human brain, utilizing a neural network model, such as deep learning, and applications of the AI system. A method of generating a text comment about content includes obtaining a content group including one or more items of content, obtaining feature information of each of the one or more items of content, determining focus content from among the one or more items of content using the obtained feature information, generating a text comment about the content group using the focus content, and displaying the generated text comment.

traducir

VEHICLE ACCIDENT IMAGE PROCESSING METHOD AND APPARATUS

NºPublicación: US2019251395A1 15/08/2019

Solicitante:

ALIBABA GROUP HOLDING LTD [KY]

CN_108399382_A

Resumen de: US2019251395A1

N vehicle accident images are obtained, where N is a natural number greater than or equal to 2. N feature vectors are obtained by inputting the vehicle accident images into a trained convolutional neural network, where the N feature vectors respectively correspond to the vehicle accident images. A distance is calculated between any two feature vectors of the N feature vectors. A determination is made that two vehicle accident images of the N vehicle accident images corresponding to the distance are abnormal when the distance is greater than a first predetermined threshold.

traducir

EFFICIENT DATA LAYOUTS FOR CONVOLUTIONAL NEURAL NETWORKS

NºPublicación: EP3523751A1 14/08/2019

Solicitante:

MAGIC LEAP INC [US]

CN_110073359_A

Resumen de: US2018096226A1

Systems and methods for efficient implementation of a convolutional layer of a convolutional neural network are disclosed. In one aspect, weight values of kernels in a kernel stack of a convolutional layer can be reordered into a tile layout with tiles of runnels. Pixel values of input activation maps of the convolutional layer can be reordered into an interleaved layout comprising a plurality of clusters of input activation map pixels. The output activation maps can be determined using the clusters of the input activation map pixels and kernels tile by tile.

traducir

Method and apparatus for monitoring electrical power consumption

NºPublicación: GB2570890A 14/08/2019

Solicitante:

GREEN RUNNING LTD [GB]

Resumen de: GB2570890A

A signal processing system for monitoring power consumption comprises: an input for receiving a power consumption signal at a first sampling frequency; an output for outputting information relating to changes in state of a plurality of devices and/or components corresponding to changes in the signal; and a processor to, for a first portion of the input signal, extract a value corresponding to each of one or more characteristics of the signal for a plurality of sampling windows, resulting in a time series of data having a data point frequency which is lower than the first sampling frequency. At least one of the characteristics corresponds to a frequency component of the signal, such as the fundamental or a harmonic. An event probability value for one or more device or component state changes is generated using a convolutional neural network (CNN) based classifier operating on the time series data, the event probability being the probability that the first portion of the signal corresponds to the change in state of the device or component. Also disclosed is a method for training the system.

traducir

System and Method for Detecting Objects in Video Sequences

NºPublicación: US2019244028A1 08/08/2019

Solicitante:

MITSUBISHI ELECTRIC RES LABORATORIES INC [US]

Resumen de: US2019244028A1

An object detector includes an input interface to accept a sequence of video frames, a memory to store a neural network trained to detect objects in the video frames, a processor to process each video frame sequentially with the neural network to detect objects in the sequence of video frames, and an output interface to output the object detection information. The neural network includes a first subnetwork, a second subnetwork, and a third subnetwork. The first subnetwork receives as an input a video frame and outputs a feature map of the video frame. The second subnetwork is a recurrent neural network that takes the feature map as an input and outputs a temporal feature map. The third subnetwork takes the temporal feature map as an input and outputs object detection information;

traducir

GESTURE RECOGNITION METHOD FOR REDUCING FALSE ALARM RATE, GESTURE RECOGNITION SYSTEM FOR REDUCING FALSE ALARM RATE, AND PERFORMING DEVICE THEREOF

NºPublicación: US2019244016A1 08/08/2019

Solicitante:

KAIKUTEK INC [TW]

US_2019244017_A1

Resumen de: US2019244016A1

A performing device of a gesture recognition system for reducing a false alarm rate executes a performing procedure of a gesture recognition method for reducing the false alarm rate. The gesture recognition system includes two neural networks. A first recognition neural network is used to classify a gesture event, and a first noise neural network is used to determine whether the sensing signal is the noise. Since the first noise neural network can determine whether the sensing signal is the noise, the gesture event may not be executed when the sensing signal is the noise. Therefore, the false alarm rate may be reduced.

traducir

RECONFIGURABLE SYSTOLIC NEURAL NETWORK ENGINE

NºPublicación: US2019244078A1 08/08/2019

Solicitante:

WESTERN DIGITAL TECH INC [US]

US_2019244085_A1

Resumen de: US2019244078A1

Some embodiments include a special-purpose hardware accelerator that can perform specialized machine learning tasks during both training and inference stages. For example, this hardware accelerator uses a systolic array having a number of data processing units (“DPUs”) that are each connected to a small number of other DPUs in a local region. Data from the many nodes of a neural network is pulsed through these DPUs with associated tags that identify where such data was originated or processed, such that each DPU has knowledge of where incoming data originated and thus is able to compute the data as specified by the architecture of the neural network. These tags enable the systolic neural network engine to perform computations during backpropagation, such that the systolic neural network engine is able to support training.

traducir

Domain Stylization Using a Neural Network Model

NºPublicación: US2019244060A1 08/08/2019

Solicitante:

NVIDIA CORP [US]

Resumen de: US2019244060A1

A style transfer neural network may be used to generate stylized synthetic images, where real images provide the style (e.g., seasons, weather, lighting) for transfer to synthetic images. The stylized synthetic images may then be used to train a recognition neural network. In turn, the trained neural network may be used to predict semantic labels for the real images, providing recognition data for the real images. Finally, the real training dataset (real images and predicted recognition data) and the synthetic training dataset are used by the style transfer neural network to generate stylized synthetic images. The training of the neural network, prediction of recognition data for the real images, and stylizing of the synthetic images may be repeated for a number of iterations. The stylization operation more closely aligns a covariate of the synthetic images to the covariate of the real images, improving accuracy of the recognition neural network.

traducir

ADJUSTING ENHANCEMENT COEFFICIENTS FOR NEURAL NETWORK ENGINE

NºPublicación: US2019244058A1 08/08/2019

Solicitante:

WESTERN DIGITAL TECH INC [US]

US_2019244085_A1

Resumen de: US2019244058A1

Some embodiments include a special-purpose hardware accelerator that can perform specialized machine learning tasks during both training and inference stages. For example, this hardware accelerator uses a systolic array having a number of data processing units (“DPUs”) that are each connected to a small number of other DPUs in a local region. Data from the many nodes of a neural network is pulsed through these DPUs with associated tags that identify where such data was originated or processed, such that each DPU has knowledge of where incoming data originated and thus is able to compute the data as specified by the architecture of the neural network. These tags enable the systolic neural network engine to perform computations during backpropagation, such that the systolic neural network engine is able to support training.

traducir

SYSTOLIC NEURAL NETWORK ENGINE WITH CROSSOVER CONNECTION OPTIMIZATION

NºPublicación: US2019244082A1 08/08/2019

Solicitante:

WESTERN DIGITAL TECH INC [US]

US_2019244081_A1

Resumen de: US2019244082A1

A method of computer processing is disclosed comprising receiving a data packet at a processing node of a neural network, performing a calculation of the data packet at the processing node to create a processed data packet, attaching a tag to the processed data packet, transmitting the processed data packet from the processing node to a receiving node during a systolic pulse, receiving the processed data packet at the receiving node, performing a clockwise convolution on the processed data packet and a counter clockwise convolution on the processed data packet, performing an adding function and backpropagating results of the performed sigmoid function to each of the processing nodes that originally processed the data packet.

traducir

MACHINE LEARNING DRIVEN DATA MANAGEMENT

NºPublicación: US2019244094A1 08/08/2019

Solicitante:

SAP SE [DE]

Resumen de: US2019244094A1

A system for machine learning driven data management is provided. In some implementations, the system performs operations including receiving, by a neural network, first and second textual data associated with a first item and a second item. The operations further include converting, by the neural network, the first and second textual data to a first vector and a second vector. The operations further include determining, by the neural network, whether the first item and the second item satisfy, based on a comparison of the first vector with the second vector, a similarity threshold. The operations further include selecting, by the neural network and in response to satisfaction of the similarity threshold, one of the first item and the second item, the selecting based on a selection criteria. The operations further include providing, by the neural network, a recommendation on a user interface regarding the selected first item or second item.

traducir

SYSTOLIC NEURAL NETWORK ENGINE CAPABLE OF FORWARD PROPAGATION

NºPublicación: US2019244077A1 08/08/2019

Solicitante:

WESTERN DIGITAL TECH INC [US]

US_2019244081_A1

Resumen de: US2019244077A1

A method of computer processing is disclosed comprising receiving a data packet at a processing node of a neural network, performing a calculation of the data packet at the processing node to create a processed data packet, attaching a tag to the processed data packet, transmitting the processed data packet from the processing node to a receiving node during a systolic pulse, receiving the processed data packet at the receiving node, performing a clockwise convolution on the processed data packet and a counter clockwise convolution on the processed data packet, performing an adding function and backpropagating results of the performed sigmoid function to each of the processing nodes that originally processed the data packet.

traducir

GESTURE RECOGNITION METHOD AND GESTURE RECOGNITION SYSTEM

Nº publicación: US2019242974A1 08/08/2019

Solicitante:

KAIKUTEK INC [TW]

US_2019244017_A1

Resumen de: US2019242974A1

A gesture recognition system executes a gesture recognition method. The gesture recognition method includes steps of: receiving a training signal; selecting one of the sensing frames of the sensing signal; generating a sensing map; selecting a cell having the max-amplitude; determining a frame amplitude, a frame phase, and a frame range of the selected one of the sensing frames; setting the frame amplitudes, the frame phases, and the frame ranges of all of the sensing frames to input data of a neural network to classify a gesture event. The present invention just uses a few data to be the input data of the neural network. Therefore, the neural network may not require high computational complexity, the gesture recognition system may decrease the calculation load of the processing unit, and the gesture recognition function may not influence a normal operation of a smart device.

traducir

Página1 de 10 nextPage por página

punteroimgVolver