| DTrOCR: Decoder-only Transformer for Optical Character Recognition | ✓ Link | 2.38 | | DTrOCR 105M | 2023-08-30 |
| Rethinking Text Line Recognition Models | | 2.75 | | Self-Attention + CTC + language model | 2021-04-15 |
| TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models | ✓ Link | 2.89 | | TrOCR-large 558M | 2021-09-21 |
| Rethinking Text Line Recognition Models | | 2.96 | | Transformer + CNN | 2021-04-15 |
| TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models | ✓ Link | 3.42 | | TrOCR-base 334M | 2021-09-21 |
| TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models | ✓ Link | 4.22 | | TrOCR-small 62M | 2021-09-21 |
| End-to-end Handwritten Paragraph Text Recognition Using a Vertical Attention Network | ✓ Link | 4.32 | 16.24 | VAN | 2020-12-07 |
| Pay Attention to What You Read: Non-recurrent Handwritten Text-Line Recognition | | 4.67 | | Transformer w/ CNN (+synth) | 2020-05-26 |
| HTR-VT: Handwritten Text Recognition with Vision Transformer | ✓ Link | 4.7 | 14.9 | HTR-VT(line-level) | 2024-09-13 |
| Easter2.0: Improving convolutional models for handwritten text recognition | ✓ Link | 6.21 | | Easter2.0 | 2022-05-30 |
| Full Page Handwriting Recognition via Image to Sequence Extraction | ✓ Link | 6.3 | | FPHR+Aug Paragraph Level (~145 dpi) | 2021-03-11 |
| Decoupled Attention Network for Text Recognition | ✓ Link | 6.4 | 19.6 | Decouple Attention Network | 2019-12-21 |
| Start, Follow, Read: End-to-End Full-Page Handwriting Recognition | ✓ Link | 6.4 | 23.2 | Start, Follow, Read | 2018-09-01 |
| Full Page Handwriting Recognition via Image to Sequence Extraction | ✓ Link | 6.5 | | FPHR+Aug Line Level (~145 dpi) | 2021-03-11 |
| No Padding Please: Efficient Neural Handwriting Recognition | ✓ Link | 6.6 | 15.9 | Leaky LP Cell | 2019-02-28 |
| Full Page Handwriting Recognition via Image to Sequence Extraction | ✓ Link | 6.7 | | FPHR Paragraph Level (~145 dpi) | 2021-03-11 |
| Pay Attention to What You Read: Non-recurrent Handwritten Text-Line Recognition | | 7.62 | | Transformer w/ CNN | 2020-05-26 |