File size: 3,062 Bytes
c661fa7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8e747c4
c661fa7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0a807c4
c661fa7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
---
license: apache-2.0
---
# DiffusionEdge Model Weights

**Important Notice: These model weights are NOT owned by the uploader of this repository. This is a redistribution of the original DiffusionEdge model weights for convenience.**

## Model Overview

This repository contains the pre-trained model weights for DiffusionEdge, a diffusion probabilistic model for crisp edge detection. These weights are redistributed here to support the MooneyMaker project, which uses DiffusionEdge for automatic creation of Mooney images.

## Original Work Attribution

- **Original Repository**: [GuHuangAI/DiffusionEdge](https://github.com/GuHuangAI/DiffusionEdge)
- **Original Authors**: Yunfan Ye, Yuhang Huang, Renjiao Yi, Zhiping Cai, Kai Xu
- **Paper**: ["DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection"](https://arxiv.org/abs/2401.02032)

## License and Legal

These model weights are redistributed under the same LICENSE (Apache-2.0) as the original DiffusionEdge repository. The original code and model weights were made available by the authors for research and development purposes. 

**Users of these weights must comply with all terms and conditions of the original license and repository.**

For any legal questions regarding the use of these weights, please refer to the original repository or contact the original authors.

## Model Details

## Usage in MooneyMaker

These weights are used in the MooneyMaker project for edge detection as part of the Mooney image generation process. The edge detection output helps to select Mooney images whose edge maps are either similar or dissimilar to those of the template image.
If you use the MooneyMaker package, the weights should be downloaded automatically. If not please place them in `mooney_maker/DiffusionEdge`.

## Original Citation

If you use these weights in your research or project, please cite the original paper:

```bibtex
@inproceedings{ye2024diffusionedge,
      title={DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection}, 
      author={Yunfan Ye and Kai Xu and Yuhang Huang and Renjiao Yi and Zhiping Cai},
      year={2024},
      booktitle={AAAI}
}
```

## Files in this Repository

- `DiffusionEdgeBSDS.pt` - Model weights trained on [BSDS500 dataset](https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/bsds/)

## Contact and Support

For questions about the model architecture, training, or original implementation:
- **Original Authors**: Contact information available in the [original repository](https://github.com/GuHuangAI/DiffusionEdge)

For questions specific to the MooneyMaker integration:
- **MooneyMaker Project**: [Link to repo]
- [Lars C. Reining](mailto:[email protected])

## Disclaimer

- All rights and ownership of the weights remain with the original authors
- For the most up-to-date weights and code, always refer to the original repository

## Acknowledgments

We thank the original authors (Yunfan Ye, Yuhang Huang, Renjiao Yi, Zhiping Cai, and Kai Xu) for making their work publicly available.

---