Network Learning Techniques (e.g., Back Propagation) Patents (Class 382/157)
  • Patent number: 11410000
    Abstract: A computer-implemented method is provided. The computer-implemented method includes classifying an image using a classification model having a residual network. Classifying the image using the classification model includes inputting an input image into the residual network having N number of residual blocks sequentially connected, N?2, (N?1) number of pooling layers respectively between two adjacent residual blocks of the N number of residual blocks, and (N?1) number of convolutional layers respectively connected to first to (N?1)-th residual blocks of the N number of residual blocks; processing outputs from the first to the (N?1)-th residual blocks of the N number of residual blocks respectively through the (N?1) number of convolutional layers; vectorizing outputs respectively from the (N?1) number of convolutional layers to generate (N?1) number of vectorized outputs; vectorizing an output from a last residual block of the N number of residual blocks to generate a last vectorized output.
    Type: Grant
    Filed: August 8, 2019
    Date of Patent: August 9, 2022
    Assignees: BEIJING BOE HEALTH TECHNOLOGY CO., LTD., BOE Technology Group Co., Ltd.
    Inventor: Xinyue Hu
  • Patent number: 11394980
    Abstract: A method of preprocessing, prior to encoding with an external encoder, image data using a preprocessing network comprising a set of inter-connected learnable weights is provided. At the preprocessing network, image data from one or more images is received. The image data is processed using the preprocessing network to generate an output pixel representation for encoding with the external encoder. The preprocessing network is configured to take as an input display configuration data representing one or more display settings of a display device operable to receive encoded pixel representations from the external encoder. The weights of the preprocessing network are dependent upon the one or more display settings of the display device.
    Type: Grant
    Filed: September 30, 2020
    Date of Patent: July 19, 2022
    Assignee: iSize Limited
    Inventors: Ioannis Andreopoulos, Srdjan Grce
  • Patent number: 11393235
    Abstract: Disclosed are computer-implemented methods, non-transitory computer-readable media, and systems for identity document face image quality recognition. One computer-implemented method includes pairing, for each user of a plurality of users and to form a pair of face images, an identity document (ID) face image and a live face image. For each pair of face images and based on a face similarity between the ID face image and the live face image, a similarity score for the ID face image is generated. Based on ID face images and similarity scores corresponding to the ID face images, a model for ID face image quality recognition is trained.
    Type: Grant
    Filed: June 25, 2021
    Date of Patent: July 19, 2022
    Assignee: ALIPAY LABS (SINGAPORE) PTE. Ltd.
    Inventor: Jianshu Li
  • Patent number: 11386307
    Abstract: A machine vision system comprising receiving means configured to receive image data indicative of an object to be classified where there is provided processing means with an initial neural network, the processing means configured to determine a differential equation describing the initial neural network algorithm based on the neural network parameters, and to determine a solution to the differential equation in the form of a series expansion; and to convert the series expansion to a finite series expansion by limiting the number of terms in the series expansion to a finite number; and to determine the output classification in dependence on the finite series expansion.
    Type: Grant
    Filed: September 25, 2018
    Date of Patent: July 12, 2022
    Assignee: NISSAN MOTOR CO., LTD.
    Inventors: Andrew Batchelor, Garry Jones, Yoshinori Sato
  • Patent number: 11386582
    Abstract: A process for reducing time of transmission for single-band, multiple-band or hyperspectral imagery using Machine Learning based compression is disclosed. The process uses Machine Learning to compress single-band, multiple-band and hyperspectral imagery, thereby decreasing the needed bandwidth and storage-capacity requirements for efficient transmission and data storage. The reduced file size for transmission accelerate the communications and reduces the transmission time. This enhances communications systems where there is a greater need for on or near real-time transmission, such as mission critical applications in national security, aerospace and natural resources.
    Type: Grant
    Filed: February 4, 2020
    Date of Patent: July 12, 2022
    Assignee: MLVX Technologies
    Inventors: Migel Dileepa Tissera, Francis George Doumet
  • Patent number: 11373060
    Abstract: A training method for video stabilization and an image processing device using the same are proposed. The method includes the following steps. An input video including low dynamic range (LDR) images is received. The LDR images are converted to high dynamic range (HDR) images by using a first neural network. A second neural network for video stabilization is trained to generate stabilized HDR images in a time-dependent manner.
    Type: Grant
    Filed: May 25, 2020
    Date of Patent: June 28, 2022
    Assignee: Novatek Microelectronics Corp.
    Inventors: Jen-Huan Hu, Wei-Ting Chen, Yu-Che Hsiao, Shih-Hsiang Lin, Po-Chin Hu, Yu-Tsung Hu, Pei-Yin Chen
  • Patent number: 11366987
    Abstract: A computer-implemented method of determining an explainability mask for classification of an input image by a trained neural network. The trained neural network is configured to determine the classification and classification score of the input image by determining a latent representation of the input image at an internal layer of the trained neural network. The method includes accessing the trained neural network, obtaining the input image and the latent representation thereof and initializing a mask for indicating modifications to the latent representation. The mask is updated by iteratively adjusting values of the mask to optimize an objective function, comprising i) a modification component indicating a degree of modifications indicated by the mask, and ii) a classification score component, determined by applying the indicated modifications to the latent representation and determining the classification score thereof. The mask is scaled to a spatial resolution of the input image and output.
    Type: Grant
    Filed: December 29, 2020
    Date of Patent: June 21, 2022
    Assignee: Robert Bosch GmbH
    Inventor: Andres Mauricio Munoz Delgado
  • Patent number: 11354548
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing images using recurrent attention. One of the methods includes determining a location in the first image; extracting a glimpse from the first image using the location; generating a glimpse representation of the extracted glimpse; processing the glimpse representation using a recurrent neural network to update a current internal state of the recurrent neural network to generate a new internal state; processing the new internal state to select a location in a next image in the image sequence after the first image; and processing the new internal state to select an action from a predetermined set of possible actions.
    Type: Grant
    Filed: July 13, 2020
    Date of Patent: June 7, 2022
    Assignee: DeepMind Technologies Limited
    Inventors: Volodymyr Mnih, Koray Kavukcuoglu
  • Patent number: 11354549
    Abstract: This disclosure relates generally to a system and method to identify various products on a plurality of images of various shelves of a retail store to facilitate compliance with respect to planograms. Planogram is a visual plan, which designates the placement of products on shelves and merchandising display fixtures of a retail store. Planograms are used to create consistency between store locations, to provide proper shelf space allocation, to improve visual merchandising appeal, and to create product-pairing suggestions. There are a few assumptions considering one instance per product class is available beforehand and the physical dimension of each product template is available in some suitable unit of length. In case of absence of physical dimension of the products, a context information of the retail store will be used. The context information is that the products of similar shapes or classes are arranged together in the shelves for consumers' convenience.
    Type: Grant
    Filed: July 14, 2020
    Date of Patent: June 7, 2022
    Assignee: Tata Consultancy Services Limited
    Inventors: Avishek Kumar Shaw, Rajashree Ramakrishnan, Shilpa Yadukumar Rao, Pranoy Hari, Dipti Prasad Mukherjee, Bikash Santra
  • Patent number: 11341364
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training an action selection neural network that is used to control a robotic agent interacting with a real-world environment.
    Type: Grant
    Filed: September 20, 2018
    Date of Patent: May 24, 2022
    Assignee: Google LLC
    Inventors: Konstantinos Bousmalis, Alexander Irpan, Paul Wohlhart, Yunfei Bai, Mrinal Kalakrishnan, Julian Ibarz, Sergey Vladimir Levine, Kurt Konolige, Vincent O. Vanhoucke, Matthew Laurance Kelcey
  • Patent number: 11341757
    Abstract: Systems and methods for generating text corpora comprising realistic optical character recognition (OCR) errors and training language models using the text corpora are provided. An example method comprises: generating, by a computer system, an initial set of images based on an input text corpus comprising text; overlaying, by the computer system, one or more simulated defects over the initial set of images to generate an augmented set of images; generating an output text corpus based on the augmented set of image; and training, using the output text corpus, a language model for optical character recognition.
    Type: Grant
    Filed: April 4, 2019
    Date of Patent: May 24, 2022
    Assignee: ABBYY Development Inc.
    Inventor: Ivan Germanovich Zagaynov
  • Patent number: 11341368
    Abstract: Methods and systems for advanced and augmented training of deep neural networks (DNNs) using synthetic data and innovative generative networks. A method includes training a DNN using synthetic data, training a plurality of DNNs using context data, associating features of the DNNs trained using context data with features of the DNN trained with synthetic data, and generating an augmented DNN using the associated features.
    Type: Grant
    Filed: April 7, 2017
    Date of Patent: May 24, 2022
    Assignee: Intel Corporation
    Inventors: Anbang Yao, Shandong Wang, Wenhua Cheng, Dongqi Cai, Libin Wang, Lin Xu, Ping Hu, Yiwen Guo, Liu Yang, Yuqing Hou, Zhou Su, Yurong Chen
  • Patent number: 11334763
    Abstract: An image processing method includes: inputting a to-be-processed image into a neural network; and forming discrete feature data of the to-be-processed image via the neural network, where the neural network is trained based on guidance information, and during the training process, the neural network is taken as a student neural network; the guidance information includes: a difference between discrete feature data formed by a teacher neural network for an image sample and discrete feature data formed by the student neural network for the image sample.
    Type: Grant
    Filed: December 2, 2019
    Date of Patent: May 17, 2022
    Assignee: BEIJING SENSETIME TECHNOLOGY DEVELOPMENT CO., LTD.
    Inventors: Yi Wei, Hongwei Qin
  • Patent number: 11328173
    Abstract: A temporal propagation network (TPN) system learns the affinity matrix for video image processing tasks. An affinity matrix is a generic matrix that defines the similarity of two points in space. The TPN system includes a guidance neural network model and a temporal propagation module and is trained for a particular computer vision task to propagate visual properties from a key-frame represented by dense data (color), to another frame that is represented by coarse data (grey-scale). The guidance neural network model generates an affinity matrix referred to as a global transformation matrix from task-specific data for the key-frame and the other frame. The temporal propagation module applies the global transformation matrix to the key-frame property data to produce propagated property data (color) for the other frame. For example, the TPN system may be used to colorize several frames of greyscale video using a single manually colorized key-frame.
    Type: Grant
    Filed: October 27, 2020
    Date of Patent: May 10, 2022
    Assignee: NVIDIA Corporation
    Inventors: Sifei Liu, Shalini De Mello, Jinwei Gu, Varun Jampani, Jan Kautz
  • Patent number: 11321587
    Abstract: A system and a method can receive a first dataset having a first label and a first context. The system and the method can also generate, at the trained deep neural network, a second dataset having the first label and a second context according to a mapping, wherein a first mapping of the plurality of mapping comprises one or more weights of the trained deep neural network that maps data having the first label and the first context to data having a second label and the first context and a second mapping of the plurality of mapping comprises one or more weights of the trained deep neural network that maps data having a second label and the first context to data having the second label and the second context, wherein the second context is different from the first context and the second label is different from the first label.
    Type: Grant
    Filed: January 30, 2020
    Date of Patent: May 3, 2022
    Assignee: Ford Global Technologies, LLC
    Inventors: Akhil Perincherry, Christopher Cruise
  • Patent number: 11314989
    Abstract: A system for training a generative model and a discriminative model. The generative model generates synthetic instances from latent feature vectors by generating an intermediate representation from the latent feature vector and generating the synthetic instance from the intermediate representation. The discriminative model determines multiple discriminator scores for multiple parts of an input instance, indicating whether the part is from a synthetic instance or an actual instance. The generative model is trained by backpropagation. During the backpropagation, partial derivatives of the loss with respect to entries of the intermediate representation are updated based on a discriminator score for a part of the synthetic instance, wherein the part of the synthetic instance is generated based at least in part on the entry of the intermediate representation, and wherein the partial derivative is decreased in value if the discriminator score indicates an actual instance.
    Type: Grant
    Filed: July 9, 2020
    Date of Patent: April 26, 2022
    Assignee: Robert Bosch GmbH
    Inventor: Andres Mauricio Munoz Delgado
  • Patent number: 11301715
    Abstract: A system and method for inserting a composited image or otherwise generated graphic into a selected video by way of a programmatic process. According to some embodiments, a system may comprise an Automated Placement Opportunity Identification (APOI) engine, a Placement Insertion Interface (PII) engine, a preview system, and an automated compositing service. The system finalizes a graphic composite into a video and provides a user with a preview for final export or further manipulation.
    Type: Grant
    Filed: August 3, 2020
    Date of Patent: April 12, 2022
    Assignee: Triple Lift, Inc.
    Inventors: Shaun T. Zacharia, Samuel Benjamin Shapiro, Alexander Prokofiev, Luis Manuel Bracamontes Hernandez
  • Patent number: 11288546
    Abstract: Provided is an apparatus for training a facial-locality super resolution deep neural network, the apparatus including a generator configured to receive a low-resolution image and convert the received low-resolution image into a fake high-resolution image similar to an original high-resolution image, a discriminator configured to compare the fake high-resolution image output from the generator with the original high-resolution image to determine authenticity, and a facial-locality loss term configured to calculate a loss that is to be minimized by the generator according to the authenticity output from the discriminator, wherein the generator is an artificial neural network learning model that learns while adjusting a weight to minimize the loss, and the facial-locality loss term calculates the loss of the generator by reflecting pixel information about a feature region of a face.
    Type: Grant
    Filed: May 31, 2019
    Date of Patent: March 29, 2022
    Assignees: Seoul National University R&DB Foundation, hodooAI Lab Inc.
    Inventors: Jungwoo Lee, Kihun Kim
  • Patent number: 11282131
    Abstract: Disclosed herein are systems, methods, and computer-readable storage devices for securely storing, at an electronic device, payment information associated with a payment account, detecting, at the electronic device, a payment operation associated with an application program. In response to the detected payment operation, the method includes determining that an input corresponding to an authorization of a payment transaction has not been locally received at the electronic device within a time period, presenting, in response to the determination, a stimulus indicating that the input corresponding to an authorization of a payment transaction has not been locally received, detecting subsequent to the stimulus presentation an authorization input, the authorization input corresponding to an authorization of a payment transaction, enabling, in response to detecting the authorization input, the payment information to be retrieved and releasing the payment information to the application program.
    Type: Grant
    Filed: September 17, 2019
    Date of Patent: March 22, 2022
    Assignee: Monticello Enterprises LLC
    Inventors: Thomas M. Isaacson, Ryan Connell Durham
  • Patent number: 11275966
    Abstract: A calculation method using pixel-channel shuffle convolutional neural network is provided. In the method, an operating system receives original input data. The original input data is pre-processed by a pixel shuffle process to be separated into multiple groups in order to minimize dimension of the data. The multiple groups of data are then processed by a channel shuffle process so as to form multiple groups of new input data selected for convolution operation. The unselected data are abandoned. Therefore, the dimension of the input data can be much effectively minimized. A multiplier-accumulator of the operating system is used to execute convolution operation using a convolution kernel and the multiple new groups of input data. Multiple output data are then produced.
    Type: Grant
    Filed: February 14, 2020
    Date of Patent: March 15, 2022
    Assignee: REALTEK SEMICONDUCTOR CORP.
    Inventors: Chun-Chang Wu, Shih-Tse Chen
  • Patent number: 11276097
    Abstract: The methods and apparatuses described herein generally relate to providing a platform for allowing a website user to select a product for purchase at a non-merchant website. For example, a commerce engine can receive a request to review product information from a non-merchant website, and can translate the request into a format that can be understood by at least one merchant server using at least one type of commerce platform. The commerce engine can send the translated requests to at least one merchant server, and the merchant servers that receive the requests can determine information about the product (e.g., remaining inventory at particular merchants, product price, and/or other product details). The merchant servers can provide this information to the commerce engine, which can send the product information to the non-merchant website. The commerce engine can also facilitate a transaction with the merchant server, based on the product information returned by the non-merchant website.
    Type: Grant
    Filed: August 21, 2019
    Date of Patent: March 15, 2022
    Assignee: PredictSpring, Inc.
    Inventors: Nitin Mangtani, Pranav Wankawala, Alex Martinovic
  • Patent number: 11263487
    Abstract: A computer-implemented technique uses a generative adversarial network (GAN) to jointly train a generator neural network (“generator”) and a discriminator neural network (“discriminator”). Unlike traditional GAN designs, the discriminator performs the dual role of: (a) determining one or more attribute values associated with an object depicted in input image fed to the discriminator; and (b) determining whether the input image fed to the discriminator is real or synthesized by the generator. Also unlike traditional GAN designs, an image classifier can make use of a model produced by the GAN's discriminator. The generator receives generator input information that includes a conditional input image and one or more conditional values that express desired characteristics of the generator output image. The discriminator receives the conditional input image in conjunction with a discriminator input image, which corresponds to either the generator output image or a real image.
    Type: Grant
    Filed: March 25, 2020
    Date of Patent: March 1, 2022
    Assignee: Microsoft Technology Licensing, LLC
    Inventors: Qun Li, Changbo Hu, Keng-hao Chang, Ruofei Zhang
  • Patent number: 11263528
    Abstract: The present disclosure provides an artificial neural network communicatively-coupled to at least one computer having one or more processors, including a plurality of neurons arranged in layers. The artificial neural network is arranged to receive a new neuron into a layer of the artificial neural network during training; the new neuron is added to the neural network when no other neuron in that layer for a selected output can learn a relationship associated with an input vector of a data set being learnt. The new neuron is updated with both the relationship which could not be learnt by any other neuron in that layer and a modified data set from a last trained neuron in that layer that contributes to the selected output of the neural network. Methods and computer-readable media are also disclosed.
    Type: Grant
    Filed: October 14, 2014
    Date of Patent: March 1, 2022
    Inventor: Bernadette Garner
  • Patent number: 11263488
    Abstract: Embodiments may provide learning and recognition of classifications using only one or a few examples of items. For example, in an embodiment, a method of computer vision processing may be implemented in a computer comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method may comprise training a neural network system implemented in the computer system to classify images into a plurality of classes using one or a few training images for each class and a plurality of associated semantic information, wherein the plurality of associated semantic information is from a plurality of sources and comprises at least some of class/object labels, textual description, or attributes, and wherein the neural network is trained by modulating the training images by sequentially applying the plurality of associated semantic information and classifying query images using the trained neural network system.
    Type: Grant
    Filed: April 13, 2020
    Date of Patent: March 1, 2022
    Assignee: International Business Machines Corporation
    Inventors: Eliyahu Schwartz, Leonid Karlinsky, Rogerio Schmidt Feris
  • Patent number: 11244377
    Abstract: Disclosed is an updated browser having an API for communicating payment data between the browser and a site or an application and a software module for processing payments of purchases and to reduce the number of user interactions needed for a purchasing process. The method includes receiving, via the user interface, an interaction by a user with an object associated with a site, the interaction indicating a user intent to make a purchase, receiving, based on the interaction and via an application programming interface, a request from the site or application for payment data in connection with the purchase and transmitting, to the site or the application and via the application programming interface, the payment data, wherein the payment data confirms the purchase or can be used to process or deliver a product associated with the purchase.
    Type: Grant
    Filed: September 17, 2019
    Date of Patent: February 8, 2022
    Assignee: MONTICELLO ENTERPRISES LLC
    Inventors: Thomas M. Isaacson, Ryan Connell Durham
  • Patent number: 11227187
    Abstract: Artificial intelligence systems are created for end users based on raw data received from the end users or obtained from any source. Training, validation and testing data is maintained securely and subject to authentication prior to use. A machine learning model is selected for providing solutions of any type or form and trained, verified and tested by an artificial intelligence engine using such data. A trained model is distributed to end users, and feedback regarding the performance of the trained model is returned to the artificial intelligence engine, which updates the model on account of such feedback before redistributing the model to the end users. When an end user provides data to an artificial intelligence engine and requests a trained model, the end user monitors progress of the training of the model, along with the performance of the model in providing quality artificial intelligence solutions, via one or more dashboards.
    Type: Grant
    Filed: May 21, 2020
    Date of Patent: January 18, 2022
    Assignee: Augustus Intelligence Inc.
    Inventor: Pascal Christian Weinberger
  • Patent number: 11222258
    Abstract: Methods, systems, and apparatus, including computer-readable media, are described for performing neural network computations using a system configured to implement a neural network on a hardware circuit. The system includes a process ID unit that receives requests to obtain data from a memory that includes memory locations that are each identified by an address. For each request, the process ID unit selects a channel controller to receive the request, provides the request to be processed by the selected channel controller, and obtains the data from memory in response to processing the request using the selected channel controller. The channel controller is one of multiple channel controllers that are configured to access any memory location of the memory. The system performs the neural network computations using the data obtained from memory and resources allocated from a shared memory of the hardware circuit.
    Type: Grant
    Filed: May 4, 2020
    Date of Patent: January 11, 2022
    Assignee: Google LLC
    Inventors: Rahul Nagarajan, Hema Hariharan
  • Patent number: 11216913
    Abstract: The present disclosure discloses a convolutional neural network processor, an image processing method and an electronic device. The method includes: receiving, by the first convolutional unit, the input image to be processed, extracting the N feature maps with different scales in the image to be processed, sending the N feature maps to the second convolutional unit, and sending the first feature map to the processing unit; fusing, by the processing unit, the received preset noise information and the first feature map, to obtain the second feature map, and sending the second feature map to the second convolutional unit; and fusing, by the second convolutional unit, the received N feature maps with the second feature map to obtain the processed image.
    Type: Grant
    Filed: April 22, 2020
    Date of Patent: January 4, 2022
    Assignee: BOE Technology Group Co., Ltd.
    Inventors: Hanwen Liu, Pablo Navarrete Michelini, Dan Zhu, Lijie Zhang
  • Patent number: 11216729
    Abstract: A recognition method includes: receiving a training voice or a training image; and extracting a plurality of voice features in the training voice, or extracting a plurality of image features in the training image; wherein when extracting the voice features, a specific number of voice parameters are generated according to the voice features, and the voice parameters are input into a deep neural network (DNN) to generate a recognition model. When extracting the image features, the specific number of image parameters are generated according to the image features, and the image parameters are input into the deep neural network to generate the recognition model.
    Type: Grant
    Filed: November 20, 2019
    Date of Patent: January 4, 2022
    Assignee: Nuvoton Technology Corporation
    Inventors: Woan-Shiuan Chien, Tzu-Lan Shen
  • Patent number: 11210394
    Abstract: In one respect, there is provided a system for training a neural network adapted for classifying one or more scripts. The system may include at least one processor and at least one memory. The memory may include program code that provides operations when executed by the at least one processor. The operations may include: reducing a dimensionality of a plurality of features representative of a file set; determining, based at least on a reduced dimensional representation of the file set, a distance between a file and the file set; and determining, based at least on the distance between the file and the file set, a classification for the file. Related methods and articles of manufacture, including computer program products, are also provided.
    Type: Grant
    Filed: October 23, 2019
    Date of Patent: December 28, 2021
    Assignee: Cylance Inc.
    Inventors: Michael Wojnowicz, Matthew Wolff, Aditya Kapoor
  • Patent number: 11200680
    Abstract: An image processing method and a related apparatus are provided. The method is applied to an image processing device, and includes: obtaining an original image, the original image including a foreground object; extracting a foreground region from the original image through a deep neural network; identifying pixels of the foreground object from the foreground region; forming a mask according to the pixels of the foreground object, the mask including mask values corresponding to the pixels of the foreground object; and extracting the foreground object from the original image according to the mask.
    Type: Grant
    Filed: November 1, 2019
    Date of Patent: December 14, 2021
    Assignee: TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED
    Inventors: Xiaolong Zhu, Kaining Huang, Jingmin Luo, Lijian Mei, Shenghui Huang, Yongsen Zheng, Yitong Wang, Haozhi Huang
  • Patent number: 11200483
    Abstract: A machine learning method based on weakly supervised learning according to an embodiment of the present invention includes extracting feature maps about a dataset given a first type of information and not given a second type of information by using a convolutional neural network (CNN), updating the CNN by back-propagating a first error value calculated as a result of performing a task corresponding to the first type of information by using a first model, and updating the CNN by back-propagating a second error value calculated as a result of performing the task corresponding to the first type of information by using a second model different from the first model, wherein the second type of information is extracted when the task corresponding to the first type of information is performed using the second model.
    Type: Grant
    Filed: December 13, 2016
    Date of Patent: December 14, 2021
    Assignee: LUNIT INC.
    Inventors: Sang Heum Hwang, Hyo Eun Kim
  • Patent number: 11200411
    Abstract: A computer model to identify a type of physical card is trained using simulated card images. The physical card may exist with various subtypes, some of which may not exist or be unavailable when the model is trained. To most robustly identify these subtypes, the training data set for the computer model includes simulated card images that are generated for the card type. The simulated card images are generated based on a semi-randomized background that varies in appearance, onto which an identifying marking of the card type is superimposed, such that the training data for the computer model includes additional randomized sample card images and ensure the model is robust to further variations in subtypes.
    Type: Grant
    Filed: October 16, 2019
    Date of Patent: December 14, 2021
    Assignee: The Toronto-Dominion Bank
    Inventors: Buturab Rizvi, Adrian Chung-Hey Ma, Ki Nam Choi, Alexandra Tsourkis
  • Patent number: 11195056
    Abstract: The technology relates to tuning a data translation block (DTB) including a generator model and a discriminator model. One or more processors may be configured to receive training data including an image in a second domain. The image in the second domain may be transformed into a first domain with a generator model. The transformed image may be processed to determine one or more outputs with one or more deep neural networks (DNNs) trained to process data in the first domain. An original objective function for the DTB may be updated based on the one or more outputs. The generator and discriminator models may be trained to satisfy the updated objective function.
    Type: Grant
    Filed: March 12, 2020
    Date of Patent: December 7, 2021
    Inventors: Alexandru Malaescu, Adrian Dorin Capata, Mihai Ciuc, Alina Sultana, Dan Filip, Liviu-Cristian Dutu
  • Patent number: 11194999
    Abstract: An integrated facial recognition method and system. The method includes: receiving a request for acquiring an integrated facial recognition service sent by a user terminal, which includes: an identifier of a user-selected model associated with facial recognition of the user, and an identifier of an operation selected by the user from candidate operations; and executing distributedly an operation selected by the user from the candidate operations on the user-selected model associated with the facial recognition of the user to obtain an operation result, and storing the operation result. The embodiment has realized completing the operations such as training a model or developing a facial recognition application, without the need of buying hardware and establishing a software environment by the user, thereby saving the development cost and improving the convenience of using the facial recognition service.
    Type: Grant
    Filed: August 14, 2018
    Date of Patent: December 7, 2021
    Assignee: Beijing Baidu Netcom Science and Technology Co., Ltd.
    Inventors: Tianhan Xu, Faen Zhang, Kai Zhou, Qian Wang, Kun Liu, Yuanhao Xiao, Dongze Xu, Jiayuan Sun, Lan Liu
  • Patent number: 11175844
    Abstract: In a deep neural network (DNN), weights are defined that represent a strength of connections between different neurons of the DNN and activations are defined that represent an output produced by a neuron after passing through an activation function of receiving an input and producing an output based on some threshold value. The weight traffic associated with a hybrid memory therefore is distinguished from the activation traffic to the hybrid memory, and one or more data structures may be dynamically allocated in the hybrid memory according to the weights and activations of the one or more data structures in the DNN. The hybrid memory includes at least a first memory and a second memory that differ according to write endurance attributes.
    Type: Grant
    Filed: May 13, 2020
    Date of Patent: November 16, 2021
    Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
    Inventors: Ashish Ranjan, Arvind Kumar, Carl Radens
  • Patent number: 11176368
    Abstract: Methods and systems for visually focused first-person neural network interpretation are disclosed. A method includes: receiving, by a computing device, an image; determining, by the computing device, feature vectors from the image; determining, by the computing device, a first padding value and a first stride value by inputting the feature vectors into a deep neural network; determining, by the computing device, a second padding value and a second stride value by inputting the feature vectors into at least one multiple regression model; determining, by the computing device, padding by averaging the first padding value and the second padding value; determining, by the computing device, stride by averaging the first stride value and the second stride value; and classifying, by the computing device, the image using a convolutional neural network using the padding and the stride.
    Type: Grant
    Filed: July 18, 2019
    Date of Patent: November 16, 2021
    Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATION
    Inventors: Aaron K. Baughman, Gopal Sarma Pingali, Maria del Pilar Belinchon Ballesta, Craig M. Trim
  • Patent number: 11175863
    Abstract: The technique of the present disclosure provides an image processing apparatus for estimating a printing result of image data to be printed with a small amount of operation after the image data is obtained. The apparatus is an image processing apparatus for estimating a printing result to be obtained by printing input image data with a printer, including: an obtaining unit that obtains the input image data; and an estimation unit that estimates the printing result based on the input image data. The estimation unit has been caused to learn scanned image data as correct data, the scanned image data being obtained by reading, with a scanner, a printing result obtained by printing predetermined image data with the printer.
    Type: Grant
    Filed: April 22, 2020
    Date of Patent: November 16, 2021
    Assignee: CANON KABUSHIKI KAISHA
    Inventor: Satoshi Ikeda
  • Patent number: 11173510
    Abstract: Systems and methods for determining the identity of a liquid within a liquid dispenser are provided. A device includes a holder having an opening that is configured to retain the liquid dispenser. A surface of the liquid dispenser is configured to affix a label identifying a liquid within the liquid dispenser. A plurality of lenses are mounted on the holder. Each lens has a respective field of view that includes a respective portion of the surface of the liquid dispenser. In addition, the device includes a plurality of fiber bundles. Each fiber bundle includes an input that is configured to receive a respective signal from a respective one of the plurality of lenses. Further, the device includes an imaging sensor that is configured to receive the respective signals from the plurality of fiber bundles and to form a plurality of images of the surface of the liquid dispenser.
    Type: Grant
    Filed: March 16, 2020
    Date of Patent: November 16, 2021
    Assignee: TWENTY TWENTY THERAPEUTICS LLC
    Inventors: Supriyo Sinha, Michael Allen, Todd Whitehurst, Dimitri Azar
  • Patent number: 11170267
    Abstract: A method, system and computer program product for region proposals are disclosed. The method includes generating a map of a video frame by calculating a plurality of pixel-level values. Each pixel-level value corresponds to a respective one of a plurality of pixels and provides an associated indication of how likely the respective one of the plurality of pixels forms part of a particular object of interest.
    Type: Grant
    Filed: June 5, 2020
    Date of Patent: November 9, 2021
    Assignee: MOTOROLA SOLUTIONS, INC.
    Inventors: Chia Ying Lee, Ying Wang, Weijuan Wu
  • Patent number: 11164300
    Abstract: Systems, methods, and computer-readable storage media for cataloguing and assessing images. This is performed by a system which receives images of an item, and identifying, within each image, the item. The system performs a structural similarity analysis of the item and for each image applies a plurality of distortions, such that for each image in the images multiple distorted images are generated. The system identifies within the distorted images at least one feature and applies a regression model to the images using the at least one feature and the structural similarity score.
    Type: Grant
    Filed: August 22, 2019
    Date of Patent: November 2, 2021
    Assignee: Walmart Apollo, LLC
    Inventors: Mani Kanteswara Garlapati, Souradip Chakraborty, Rajesh Shreedhar Bhat
  • Patent number: 11164071
    Abstract: Disclosed herein is convolutional neural network (CNN) system for generating a classification for an input image. According to an embodiment, the CNN system comprises a sequence of neural network layers configured to: derive a feature map based on at least the input image; puncture at least one selection among the feature map and a kernel by setting the value of one or more elements of a row of the at least one selection to zero according to a pattern and cyclic shifting the pattern by a predetermined interval per row to set the value of one or more elements of the rest of the rows of the at least one selection according to the cyclic shifted pattern; convolve the feature map with the kernel to generate a first convolved output; and generate the classification for the input image based on at least the first convolved output.
    Type: Grant
    Filed: June 27, 2017
    Date of Patent: November 2, 2021
    Inventors: Mostafa El-Khamy, Yoo Jin Choi, Jungwon Lee
  • Patent number: 11157775
    Abstract: A computer constructs a neural network for executing image processing, the neural network being constituted of layers, each of which includes at least one node. The neural network includes a detection layer that realizes a process for detecting an object in an image. The computer is configured to execute: a first process of obtaining setting information for constructing the neural network including setting values relating to characteristics of a boundary of the object and a shape of the object, the setting values being values for calculating hyperparameters of the detection layer; a second process of constructing the neural network on the basis of the setting information. The second process includes a process of calculating the hyperparameters of the detection layer on the basis of the setting values.
    Type: Grant
    Filed: October 31, 2019
    Date of Patent: October 26, 2021
    Assignee: HITACHI, LTD.
    Inventor: Masahiro Kageyama
  • Patent number: 11157771
    Abstract: A method for learning deep convolutional features specifically designed for correlation filter based visual tracking includes the steps of, selecting a first image from a first image patch; selecting a second image from a second image patch; forward propagating selected first image by a convolutional neural network model formula, the formula has random weights with zero mean for the parameters; forward propagating selected second image by the convolutional neural network model formula; computing correlation filter using forward propagated second image and centered correlation response; circularly correlating forward propagated first image and computed correlation filter to generate predicted response map; calculating the loss by comparing the predicted response map with desired correlation corresponding selected first image and second image and updating the parameters of the convolutional neural network model formula according to calculated loss.
    Type: Grant
    Filed: May 12, 2017
    Date of Patent: October 26, 2021
    Assignees: ASELSAN ELEKTRONIK SANAYI VE TICARET ANONIM SIRKETI, ORTA DOGU TEKNIK UNIVERSITESI
    Inventors: Erhan Gundogdu, Abdullah Aydin Alatan
  • Patent number: 11151744
    Abstract: Methods for annotating objects within image frames are disclosed. Information is obtained that represents a camera pose relative to a scene. The camera pose includes a position and a location of the camera relative to the scene. Data is obtained that represents multiple images, including a first image and a plurality of other images, being captured from different angles by the camera relative to the scene. A 3D pose of the object of interest is identified with respect to the camera pose in at least the first image. A 3D bounding region for the object of interest in the first image is defined, which indicates a volume that includes the object of interest. A location and orientation of the object of interest is determined in the other images based on the defined 3D bounding region of the object of interest and the camera pose in the other images.
    Type: Grant
    Filed: September 16, 2019
    Date of Patent: October 19, 2021
    Assignee: X Development LLC
    Inventors: Kurt Konolige, Nareshkumar Rajkumar, Stefan Hinterstoisser, Paul Wohlhart
  • Patent number: 11151829
    Abstract: The invention refers to a computer-implemented method of analyzing security documents having visible information and at least one of infrared and ultraviolet detectable information, the method comprising the steps of: receiving visible-color data of a first set of pixels of a first region of the security document in a first image of the security document and feeding the visible-color data to a convolutional neural network, CNN; receiving infrared, IR, and/or ultraviolet, UV, data of a second and/or third set of pixels of a respective second and/or third region in a respective second and/or third image of the security document and feeding the IR and/or UV data to the CNN; analyzing the visible-color data of the first set of pixels using the CNN to extract characteristics of the security document from the visible information; and analyzing the IR and/or UV data of the second and/or third set of pixels using the CNN to extract characteristics of the security document from the IR and/or UV information.
    Type: Grant
    Filed: May 23, 2019
    Date of Patent: October 19, 2021
    Assignee: IDEMIA Identity & Security Germany AG
    Inventor: Norbert Wendt
  • Patent number: 11151414
    Abstract: Methods, devices, systems and apparatus for training image enhancement models and enhancing images are provided. In one aspect, a method of training an image enhancement model includes: for each of one or more constraint features, processing a ground truth image with the constraint feature to obtain a feature image corresponding to the constraint feature, for each of the one or more feature images, using the ground truth image and the feature image to train a convolutional neural network (CNN) structure model corresponding to the feature image, determining a loss function of the image enhancement model based on the one or more CNN structure models corresponding to the one or more feature images, and establishing the image enhancement model based on the loss function. A to-be-enhanced image can be input into the established image enhancement model to obtain an enhanced image.
    Type: Grant
    Filed: May 12, 2020
    Date of Patent: October 19, 2021
    Assignee: Shanghai Neusoft Medical Technology Co., Ltd.
    Inventor: Feng Huang
  • Patent number: 11151707
    Abstract: A system for defect review and classification is disclosed. The system may include a controller, wherein the controller may be configured to receive one or more training images of a specimen. The one or more training images including a plurality of training defects. The controller may be further configured to apply a plurality of difference filters to the one or more training images, and receive a signal indicative of a classification of a difference filter effectiveness metric for at least a portion of the plurality of difference filters. The controller may be further configured to generate a deep learning network classifier based on the received classification and the attributes of the plurality of training defects. The controller may be further configured to extract convolution layer filters of the deep learning network classifier, and generate one or more difference filter recipes based on the extracted convolution layer filters.
    Type: Grant
    Filed: February 15, 2019
    Date of Patent: October 19, 2021
    Assignee: KLA Corporation
    Inventors: Santosh Bhattacharyya, Jacob George, Saravanan Paramasivam, Martin Plihal
  • Patent number: 11138474
    Abstract: A parameter training method for a convolutional neural network, CNN, for detecting items of interest visible in images by a data processor of at least one server. The method is implemented based on a plurality of training image databases. The items of interest are already annotated, the CNN being a CNN common to the plurality of training image databases and having a common core and a plurality of encoding layers, each one specific to one of the plurality of training image databases. The method is also for detecting items of interest visible in an image.
    Type: Grant
    Filed: October 2, 2019
    Date of Patent: October 5, 2021
    Assignee: IDEMIA IDENTITY & SECURITY FRANCE
    Inventors: Cécile Jourdas, Dora Csillag, Maxime Thiebaut
  • Patent number: 11138689
    Abstract: The present invention provides a method for non-linearly stretching a cropped image, including: obtaining an FOV difference table of a panoramic lens; calculating a polynomial of an approximated curve represented by the values of the FOV difference table; shooting a circular wide-angle image using the panoramic lens; cropping the circular wide-angle image to preserve the remaining circular wide-angle image under a specific viewing angle; under the specific viewing angle, resampling a plurality of points by the polynomial to obtain the corresponding spacing distance of the points; and while maintaining the ratio of the spacing distance between the plurality of points, stretching the remaining circular wide-angle image to the size of the original circular wide-angle image.
    Type: Grant
    Filed: November 26, 2018
    Date of Patent: October 5, 2021
    Assignee: QUANTA COMPUTER INC.
    Inventors: Chun-Chieh Chang, Chung-Run Liao