|
|
--- |
|
|
license: mit |
|
|
library_name: pytorch |
|
|
pipeline_tag: image-to-image |
|
|
tags: |
|
|
- document-image-restoration |
|
|
- dewarping |
|
|
- deshadowing |
|
|
- deblurring |
|
|
- binarization |
|
|
- appearance-enhancement |
|
|
- doctra |
|
|
model-index: |
|
|
- name: DocRes (Main Weights, docres.pkl) |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
# DocRes Main Weights (docres.pkl) |
|
|
|
|
|
These are the official **DocRes** (CVPR 2024) main weights (`docres.pkl`), rehosted for use in the [Doctra](https://github.com/AdemBoukhris457/Doctra) library. |
|
|
|
|
|
--- |
|
|
|
|
|
## 📖 Source |
|
|
- Original repository: [ZZZHANG-jx/DocRes](https://github.com/ZZZHANG-jx/DocRes) |
|
|
- Paper: *DocRes: Dynamic Task-Specific Prompt for Generalist Document Image Restoration* (CVPR 2024) |
|
|
|
|
|
--- |
|
|
|
|
|
## ⚖️ License |
|
|
MIT License (see LICENSE file). |
|
|
Weights are redistributed under the same terms, with attribution to the original authors. |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ Intended Use |
|
|
The `docres.pkl` weights are used with the DocRes model backbone to perform generalist document image restoration tasks, including: |
|
|
|
|
|
- 📐 Dewarping |
|
|
- 🌗 Deshadowing |
|
|
- ✨ Appearance enhancement (illumination cleanup) |
|
|
- 🔄 Deblurring |
|
|
- ⚫ Binarization |
|
|
|
|
|
These weights are integrated into the **Doctra** library to improve preprocessing and restoration of scanned or photographed documents. |
|
|
|
|
|
--- |
|
|
|
|
|
## ⚠️ Limitations |
|
|
- Performance may not always exceed highly specialized single-task models. |
|
|
- Trained on specific datasets (see [source repo](https://github.com/ZZZHANG-jx/DocRes) for details). |
|
|
- Not intended for non-document natural images. |
|
|
|
|
|
--- |
