Lightweight document layout analysis method

The invention discloses a lightweight document layout analysis method, which comprises the following steps of: acquiring and labeling a first document image and a first label representing the relative position and category of a layout element, combining the first document image and the first label i...

Full description

Saved in:
Bibliographic Details
Main Authors YANG CIHUI, CUI GUOHAO, LI QIANHONG, YANG DESHENG, XU SIYAO, CAI YUNBING
Format Patent
LanguageChinese
English
Published 11.06.2024
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a lightweight document layout analysis method, which comprises the following steps of: acquiring and labeling a first document image and a first label representing the relative position and category of a layout element, combining the first document image and the first label into a first image label pair, and adding the first image label pair into a first document layout analysis data set; preprocessing the first document layout analysis data set to obtain a second document layout analysis data set, and dividing the second document layout analysis data set into a training set, a verification set and a test set; designing a first document layout analysis model structure; initializing model parameters, and training the first document layout analysis model by using the second document layout analysis data set to obtain a second document layout analysis model; and inputting a to-be-detected image to the second document layout analysis model for analysis, and outputting a layout primitive re
Bibliography:Application Number: CN202410567230