A two-branch CNN-Transformer model with a symbol-counting auxiliary task improves handwritten math expression recognition by 0.4 to 1.5 percentage points over PosFormer on CROHME 2014, 2016, and 2019.
Learned image compression with mixed transformer-cnn architectures[C]//Proceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
MMHMER:Multi-viewer and Multi-task for Handwritten Mathematical Expression Recognition
A two-branch CNN-Transformer model with a symbol-counting auxiliary task improves handwritten math expression recognition by 0.4 to 1.5 percentage points over PosFormer on CROHME 2014, 2016, and 2019.