利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

517 条记录 (耗时 0.036 秒)

    Extracting text from an image
    5.
    发明授权
    Extracting text from an image 有权

    公开(公告)号：US12062246B2

    公开(公告)日：2024-08-13

    申请号：US17490770

    申请日：2021-09-30

    申请人： Konica Minolta Business Solutions U.S.A., Inc.

    发明人： Tim Prebble

    IPC分类号： G06V30/18 ,  G06F18/2431 ,  G06T7/11 ,  G06T7/13 ,  G06T11/00

    CPC分类号： G06V30/18 ,  G06F18/2431 ,  G06T7/11 ,  G06T7/13 ,  G06T11/00

    摘要： A method for extracting text from an input image and generating a document includes: generating an edges mask from the input image; generating an edges image that is derived from the edges mask; identifying, within the edges mask, one or more probable text areas; extracting a first set of text characters by performing a first optical character recognition (OCR) operation on each of one or more probable text portions, of the derived edges image, corresponding to each of the probable text areas; generating a modified image by erasing, from the input image, image characters corresponding to the first set of text characters extracted by the first OCR operation; and generating a document by overlaying the extracted first set of text characters on the modified image.

    Methods and systems for watermarking neural networks
    9.
    发明授权
    Methods and systems for watermarking neural networks 有权

    公开(公告)号：US12050671B2

    公开(公告)日：2024-07-30

    申请号：US17858775

    申请日：2022-07-06

    申请人： Nanyang Technological University

    发明人： Nandish Chattopadhyay ,  Anupam Chattopadhyay

    IPC分类号： G06F21/16 ,  G06F18/2431 ,  G06N3/08

    CPC分类号： G06F21/16 ,  G06F18/2431 ,  G06N3/08

    摘要： Disclosed herein is a system for watermarking a neural network, comprising memory; and at least one processor in communication with the memory; wherein the memory stores instructions for causing the at least one processor to carry out a method comprising: generating a trigger set by obtaining examples from a training set by random sampling from the training set, respective examples being associated with respective true classes of a plurality of classes; generating a set of adversarial examples by structured perturbation of the examples; generating, for each adversarial example, one or more adversarial class labels by passing the adversarial example to the neural network; and applying one or more trigger labels to each said adversarial example, wherein the one or more trigger labels are selected randomly from the plurality of classes, and wherein each trigger label is not a said true class label for the corresponding example or a said adversarial class label for the corresponding adversarial example; and storing the adversarial examples and corresponding trigger labels as the trigger set; and performing a tuning process to adjust parameters at each layer of the neural network using the trigger set, to thereby generate a watermarked neural network.

    Distance-based learning confidence model
    10.
    发明授权
    Distance-based learning confidence model 有权

    公开(公告)号：US12039443B2

    公开(公告)日：2024-07-16

    申请号：US18045722

    申请日：2022-10-11

    申请人： Google LLC

    发明人： Sercan Omer Arik ,  Chen Xing ,  Zizhao Zhang ,  Tomas Jon Pfister

    IPC分类号： G06N3/08 ,  G06F18/214 ,  G06F18/2413 ,  G06F18/2431 ,  G06N3/04

    CPC分类号： G06N3/08 ,  G06F18/2148 ,  G06F18/2413 ,  G06F18/2431 ,  G06N3/04

    摘要： A method includes receiving a training data set including a plurality of training data subsets. From two or more training data subsets in the training data set, the method includes selecting a support set of training examples and a query set of training examples. The method includes determining, using the classification model, a centroid value for each respective class. For each training example in the query set of training examples, the method includes generating, using the classification model, a query encoding, determining a class distance measure, determining a ground-truth distance, and updating parameters of the classification model. For each training example in the query set of training examples identified as being misclassified, the method further includes generating a standard deviation value, sampling a new query, and updating parameters of the confidence model based on the new query encoding.