End-to-end table detection and structure identification method and system

A table structure and structure recognition technology, applied in the field of computer vision, can solve the problem of inability to deal with the lack of dividing lines or the complete lack of wireless tables, and achieve the effect of improving robustness and versatility and reducing difficulty.

Pending Publication Date: 2021-09-24
BEIJING YIDAO BOSHI TECH
View PDF0 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method can also achieve better performance in wired tables, but it cannot deal with wireless tables with partly or completely missing dividers

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • End-to-end table detection and structure identification method and system
  • End-to-end table detection and structure identification method and system
  • End-to-end table detection and structure identification method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0081] Step 1: Image Preprocessing

[0082] This step performs a series of preprocessing operations on the input image. The image contains one or more tables. In view of the fact that most of the tables have the characteristics of compact rows and small row spacing, in order to improve the distinction between rows, this design first completes the vertical stretch transformation at this stage to increase the gap between rows. pixel distance between . The next preprocessing operation also includes normalizing the size of the image with a constant aspect ratio and adding 0 to the boundary, so that the size of the image can support the requirements of the neural network and maximize the retention of global and local feature information. During training, the image preprocessing stage also needs to complete the necessary data enhancement, such as image affine transformation (Rotation, Shear, Scale, etc.), color distortion, etc., so that the distribution of training samples is clos...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an end-to-end table detection and structure identification method and system, and relates to the field of computer vision. The method comprises the following steps: stretching an original image in the vertical direction, and carrying out dimension normalization with an invariable length-width ratio and boundary 0 supplementation to form a preprocessed image; taking an encoder-decoder model as a main body structure, determining a table area in the preprocessed image, and classifying the table area into a wired table image and a wireless table image; based on the determined table area, separating a corrected table area image only containing the table area from the preprocessed image; and for the table region image, performing table structure identification by adopting different methods according to the classified wired table image and wireless table image. Different structure recognition methods are adopted for different types of tables, and the robustness and universality of the algorithms are improved by fully combining the advantages of a convolutional neural network image segmentation algorithm, an image convolutional neural network algorithm and a traditional rule analysis method.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to an end-to-end table detection and structure recognition method and system. Background technique [0002] In reality, tables, as a way of carrying key information, widely exist in objects such as PDFs, scanned documents, and photographed pictures. Table structure recognition is an important prerequisite for many downstream tasks, such as document analysis, information extraction, and visualization. Automatic table recognition methods usually include two steps: table detection and table structure recognition. Among them, the purpose of table detection is to locate the table area in the picture, and table recognition is to identify the internal structure of the table in each area to obtain the final result. structured data. Manually extracting table content will consume a lot of manpower and time. In contrast, the automated approach will greatly improve work efficiency. [0003] T...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00G06K9/20G06K9/32G06K9/34G06K9/46G06K9/62G06T5/30G06T3/40G06N3/04G06N3/08
CPCG06T5/30G06T3/4007G06T3/4053G06N3/08G06N3/045G06F18/24
Inventor 周勃宇王勇朱军民
Owner BEIJING YIDAO BOSHI TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap