Title:
画像処理方法及び装置、電子機器並びに記憶媒体
Document Type and Number:
Japanese Patent JP7097513
Kind Code:
B2
Abstract:
An image processing method and apparatus, an electronic device, and a storage medium. The method comprises: performing recognition on an image, and determining multiple target regions in the image (S11), wherein the target regions are regions in which text to be extracted is located; determining relative position features between the respective target regions in the image (S12); determining target features of the respective target regions (S13), the target features comprising a feature of the text to be extracted; performing feature extraction on the relative position features and the target features by means of a graph convolutional neural network, so as to obtain extracted features (S14); and determining, according to the extracted features, fields corresponding to the text to be extracted (S15).
Inventors:
San Hongbin
Yue Xiaoyu
Quan Chang Hui
Lin Chen Hao
Chang Wei
Yue Xiaoyu
Quan Chang Hui
Lin Chen Hao
Chang Wei
Application Number:
JP2021538344A
Publication Date:
July 07, 2022
Filing Date:
February 28, 2020
Export Citation:
Assignee:
SHENZHEN SENSETIME TECHNOLOGY CO.,LTD
International Classes:
G06T7/00; G06T7/162
Domestic Patent References:
| JP2019109562A | ||||
| JP2019105915A | ||||
| JP2019191783A | ||||
| JP2000132639A |
Attorney, Agent or Firm:
Kenji Sugimura
Mitsutsugu Sugimura
Kiyoshi Shigeno
Mitsutsugu Sugimura
Kiyoshi Shigeno