mlsanityData to 7 Oct 2026

MathNet: A Data-Centric Approach for Printed Mathematical Expression Recognition

arXiv 2404.13667 ↗First posted 21 Apr 2024cs.CV cs.AIcited by 4

Not on the map. Why?

Abstract (arXiv)

Printed mathematical expression recognition (MER) models are usually trained and tested using LaTeX-generated mathematical expressions (MEs) as input and the LaTeX source code as ground truth. As the same ME can be generated by various different LaTeX source codes, this leads to unwanted variations in the ground truth data that bias test performance results and hinder efficient learning. In addition, the use of only one font to generate the MEs heavily limits the generalization of the reported results to realistic scenarios. We propose a data-centric approach to overcome this problem, and present convincing experimental results: Our main contribution is an enhanced LaTeX normalization to map any LaTeX ME to a canonical form. Based on this process, we developed an improved version of the benchmark dataset im2latex-100k, featuring 30 fonts instead of one. Second, we introduce the real-world dataset realFormula, with MEs extracted from papers. Third, we developed a MER model, MathNet, based on a convolutional vision transformer, with superior results on all four test sets (im2latex-100k, im2latexv2, realFormula, and InftyMDB-1), outperforming the previous state of the art by up to 88.3%.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Builds on 14 of its 36 references that are in the map. First posted on arXiv 21 Apr 2024.

Builds on

14 of 36 references.

References · 14

  1. 2017Attention Is All You Needcited by 59,932
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,016
  5. 2014On the Properties of Neural Machine Translation: Encoder-Decoder Approachescited by 1,303
  6. 2021CvT: Introducing Convolutions to Vision Transformerscited by 946
  7. 2017A Closer Look at Memorization in Deep Networkscited by 936
  8. 2023Nougat: Neural Optical Understanding for Academic Documentscited by 240
  9. 2016Image-to-Markup Generation with Coarse-to-Fine Attentioncited by 79
  10. 2022When Counting Meets HMER: Counting-Aware Network for Handwritten Mathematical Expression Recognitioncited by 35
  11. 2020Improving Attention-Based Handwritten Mathematical Expression Recognition with Scale Augmentation and Drop Attentioncited by 22
  12. 2021Handwritten Mathematical Expression Recognition via Attention Aggregation based Bi-directional Mutual Learningcited by 20
2 more
  1. 2020ConvMath: A Convolutional Sequence Network for Mathematical Expression Recognitioncited by 9
  2. 2018Teaching Machines to Code: Neural Markup Generation with Visual Attentioncited by 5

Built on by

Cited by year

Citing papers per year: 2025: 4.

4 papers.

  1. 2025SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversioncited by 51
  2. 2025Uni-MuMER: Unified Multi-Task Fine-Tuning of Vision-Language Model for Handwritten Mathematical Expression Recognitioncited by 3
  3. 2025AutoOpt: A Dataset and a Unified Framework for Automating Optimization Problem Solvingcited by 1
  4. 2025A document is worth a structured record: Principled inductive bias design for document recognitioncited by 0