Title:
DEEP LEARNING MODEL TRAINING METHOD FOR TEXT DETECTION, AND TEXT DETECTION METHOD
Document Type and Number:
WIPO Patent Application WO/2023/015939
Kind Code:
A1
Abstract:
Provided in the present disclosure are a deep learning model training method for text detection, and a text detection method, which relate to the technical field of artificial intelligence, in particular to the technical fields of computer vision and deep learning, and may be applied to scenarios such as optical character recognition (OCR). The deep learning model training method for text detection is provided. A single character segmentation sub-network outputs a single character segmentation prediction result. A text line segmentation sub-network outputs a text line segmentation prediction result. A trained deep learning model can be used to detect a text area; and predictions for single character segmentation and text line segmentation can be simultaneously realized, such that text detection can be performed by means of combining the two text segmentation modes, and thus the accuracy of text area detection can be further improved.
More Like This:
Inventors:
FAN SEN (CN)
WANG XIAOYAN (CN)
LV PENGYUAN (CN)
ZHANG CHENGQUAN (CN)
YAO KUN (CN)
WANG XIAOYAN (CN)
LV PENGYUAN (CN)
ZHANG CHENGQUAN (CN)
YAO KUN (CN)
Application Number:
PCT/CN2022/088280
Publication Date:
February 16, 2023
Filing Date:
April 21, 2022
Export Citation:
Assignee:
BEIJING BAIDU NETCOM SCI & TECH CO LTD (CN)
International Classes:
G06K9/00
Foreign References:
| CN113642583A | 2021-11-12 | |||
| CN110766002A | 2020-02-07 | |||
| CN111639646A | 2020-09-08 | |||
| CN109002463A | 2018-12-14 | |||
| US20190180154A1 | 2019-06-13 | |||
| CN109697232A | 2019-04-30 |
Attorney, Agent or Firm:
PATENTSINO IP FIRM (CN)
Download PDF: