P

PP DocLayout Plus L

Developed by PaddlePaddle
PP-DocLayout_plus-L is a high-precision document layout area positioning model, trained based on the RT-DETR-L architecture, and supports the detection of 20 common document elements.
Downloads 1,308
Release Time : 6/6/2025

Model Overview

This model is specifically designed for the layout analysis of document images and can accurately locate various elements such as titles, paragraphs, tables, and formulas in documents. It is suitable for processing documents mixed with Chinese and English.

Model Features

Multi-category detection
Supports the detection of 20 document elements, including text, titles, tables, formulas, etc.
High precision
Achieves an mAP(0.5) index of 83.2% on the self-built dataset
Wide applicability
The training data covers various document types such as papers, PPTs, contracts, and ancient books

Model Capabilities

Document layout analysis
Table detection
Formula detection
Title recognition
Text area positioning

Use Cases

Document processing
Academic paper analysis
Automatically identify the structure of titles, abstracts, references, etc. in papers
Can generate structured paper content
Contract analysis
Locate the key clauses and signature areas in the contract
Assist in the contract review process
Education
Exam paper analysis
Identify the question and answer areas in the exam paper
Assist in the automatic grading system
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase