Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
**<span style="color: #A6351B;">Warning: this project is under a research/academic usage only license, preventing any usage for commercial purposes without permission.</span>**
|
| 2 |
+
|
| 3 |
+
This repository contains official trained model weights from the paper:
|
| 4 |
+
Denis Coquenet, "Meta-DAN: towards an efficient prediction strategy for page-level handwritten text recognition", Pattern Recognition, 2026.
|
| 5 |
+
|
| 6 |
+
The paper is available [here](https://doi.org/10.1016/j.patcog.2026.113373).
|
| 7 |
+
The source code is available [here](https://github.com/FactoDeepLearning/META-DAN).
|
| 8 |
+
|
| 9 |
+
Weights are given for the following models: DAN (TPAMI 2023), Faster DAN (ICDAR 2023), MT-DAN (PR 2026), W-DAN (PR 2026) and Meta-DAN (PR 2026).
|
| 10 |
+
Pre-trained weights at line level are also provided.
|
| 11 |
+
|
| 12 |
+
Each models is either trained on a single dataset (BRESSAY, CASIA2, Eparchos, Esaposalles, IAM, MAURDOR, READ2016, RIMES, ScribbleLens) or on a set of latin languages (IAM+BRESSAY+READ2016+Espossales+ScribbleLens).
|
| 13 |
+
|