Texteller
TexTeller is an end-to-end formula recognition model based on the ViT architecture, capable of directly converting mathematical formulas in natural scene images into LaTeX format.
Downloads 19
Release Time : 8/14/2024
Model Overview
TexTeller is an image-to-text model specifically designed to recognize mathematical formulas in natural scene images and convert them into LaTeX format.
Model Features
Large-scale training data
The training dataset has been expanded to 7.5 million, approximately 15 times that of TexTeller 1.0, with significantly improved data quality.
Outstanding performance
Demonstrates outstanding performance in test sets, particularly excelling in rare symbol recognition, complex multi-line formulas, and matrix parsing.
Strong generalization ability
Compared to LaTeX-OCR, which uses only about 100,000 data points, it exhibits superior generalization ability and recognition accuracy.
Model Capabilities
Mathematical formula recognition in images
LaTeX format conversion
Complex formula parsing
Rare symbol recognition
Use Cases
Academic research
Paper formula recognition
Convert mathematical formula images in academic papers into LaTeX format for easy editing and citation.
Improves formula editing efficiency and reduces manual input errors.
Education
Teaching material creation
Convert handwritten or printed mathematical formulas into LaTeX format for creating teaching materials.
Simplifies the process of creating teaching materials and improves the accuracy and aesthetics of formulas.
Featured Recommended AI Models
Š 2025AIbase