Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
task_categories:
|
4 |
+
- text-to-image
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
---
|
8 |
+
# HumanRefiner
|
9 |
+
|
10 |
+
![Overview](https://raw.githubusercontent.com/Enderfga/HumanRefiner/main/framework.jpg)
|
11 |
+
|
12 |
+
## Introduction
|
13 |
+
|
14 |
+
Welcome to the official code repository for the [paper](https://arxiv.org/abs/2407.06937) "HumanRefiner: Benchmarking Abnormal Human Generation and Refining with Coarse-to-fine Pose-Reversible Guidance."
|
15 |
+
|
16 |
+
In this project, we introduce AbHuman, the first large-scale benchmark focused on anatomical anomalies. The benchmark consists of 56K synthesized human images, each annotated with 147K human anomalies in 18 different categories. Based on this, we developed HumanRefiner, a novel plug-and-play method for coarse-to-fine refinement of human anomalies.
|
17 |
+
|
18 |
+
## Data
|
19 |
+
|
20 |
+
Our data is available on 🤗 [Hugging Face](https://huggingface.co/datasets/Enderfga/HumanRefiner).
|
21 |
+
|
22 |
+
To download the dataset, use the following commands:
|
23 |
+
|
24 |
+
```bash
|
25 |
+
git lfs install
|
26 |
+
git clone https://huggingface.co/datasets/Enderfga/HumanRefiner
|
27 |
+
```
|
28 |
+
|
29 |
+
After cloning, extract the train and validation datasets:
|
30 |
+
|
31 |
+
```bash
|
32 |
+
tar -xzf train.zip
|
33 |
+
tar -xzf val.zip
|
34 |
+
```
|
35 |
+
## Annotation Examples
|
36 |
+
Below is a detailed illustration of class definitions with visualized examples from the AbHuman dataset:
|
37 |
+
![](https://raw.githubusercontent.com/Enderfga/HumanRefiner/main/example.jpg)
|
38 |
+
|
39 |
+
## Cite
|
40 |
+
|
41 |
+
If you use our work in your research, please cite our paper:
|
42 |
+
|
43 |
+
```bibtex
|
44 |
+
@misc{fang2024humanrefinerbenchmarkingabnormalhuman,
|
45 |
+
title={HumanRefiner: Benchmarking Abnormal Human Generation and Refining with Coarse-to-fine Pose-Reversible Guidance},
|
46 |
+
author={Guian Fang and Wenbiao Yan and Yuanfan Guo and Jianhua Han and Zutao Jiang and Hang Xu and Shengcai Liao and Xiaodan Liang},
|
47 |
+
year={2024},
|
48 |
+
eprint={2407.06937},
|
49 |
+
archivePrefix={arXiv},
|
50 |
+
primaryClass={cs.CV},
|
51 |
+
url={https://arxiv.org/abs/2407.06937},
|
52 |
+
}
|
53 |
+
```
|
54 |
+
|
55 |
+
## Contact
|
56 |
+
|
57 |
+
If you have any questions or suggestions, please contact us:
|
58 |
+
|
59 |
+
- Email: [email protected]
|
60 |
+
- GitHub Issues: [HumanRefiner Issues](https://github.com/Enderfga/HumanRefiner/issues)
|
61 |
+
|
62 |
+
Thank you for your support!
|