https://github.com/yumingj/GroupDiff

https://github.com/yumingj/GroupDiff

Last synced: 5 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/yumingj/GroupDiff
• Owner: yumingj
• Created: 2024-07-14T16:22:04.000Z (7 months ago)
• Default Branch: main
• Last Pushed: 2024-07-14T16:34:36.000Z (7 months ago)
• Last Synced: 2024-09-28T06:02:44.499Z (4 months ago)
• Language: Python
• Size: 21.5 KB
• Stars: 6
• Watchers: 4
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-diffusion-categorized - [Code

README

        


GroupDiff: Diffusion-based Group Portrait Editing




## Dataset

Our dataset is developed based on NUS LV Multiple-Human Parsing Dataset v2.0. Please download the source data from the [link](https://drive.google.com/file/d/1YVBGMru0dlwB8zu1OoErOazZoc8ISSJn/view?usp=sharing).

We use [MMPose](https://github.com/open-mmlab/mmpose) to estimate the pose using the "vitpose_h" model. You can download the pose estimation results from this [link](https://drive.google.com/file/d/1_ivJ5jTv0p-gdcZ8XLvTix_ymg7KOJTL/view?usp=sharing).

After downloading the dataset, unzip the file and put them under the dataset folder with the following structure:

```

./LV-MHP-v2

├── train

    ├── images

    ├── parsing_annos

    └── pose_estimation

└── shhq_dataset

    ├── images

    ├── parsing_annos

    └── pose_estimation

```

Based on the preprocessed data, we propose a comprehensive training data generation engine to synthesize paired data. The data generation codes can be found [here](https://github.com/yumingj/GroupDiff/blob/main/data/group_diff_data_gen.py).