BetterDFLExtractor is designed as a more efficient and powerful alternative to the standard extractor in DeepFaceLab. It aims to improve the face extraction process used in deepfake creation, focusing on enhanced performance and accuracy.
This project incorporates developments from the 3DDFA_V2
project. Special thanks to Jianzhu Guo, Xiangyu Zhu, Yang Yang, Fan Yang, Zhen Lei, and Stan Z Li for their significant contributions to the field of 3D face alignment.
Relevant references:
- Guo, J., Zhu, X., Yang, Y., Yang, F., Lei, Z., & Li, S. Z. (2020). Towards Fast, Accurate and Stable 3D Dense Face Alignment. Proceedings of the European Conference on Computer Vision (ECCV).
- Guo, J., Zhu, X., & Lei, Z. (2018). 3DDFA. Retrieved from 3DDFA GitHub Repository
This project is licensed under the MIT License.
- Clone this repository into your DeepFaceLab directory:
git clone https://github.com/Jackmjedi/BetterDFLExtractor .
in the DeepFaceLab directory.- For Windows:
_internal/DeepFaceLab
- For Linux:
DeepFaceLab_Linux/DeepFaceLab
- For Windows:
- Use Anaconda to create an environment using the
environment.yaml
file provided.conda env create -f environment.yaml
- Run
python caller.py
in the DeepFaceLab folder to launch the extractor. - Select video, then extract. After selecting your desired cluster and exporting, it will be placed in workspace/faces.
Please note that BetterDFLExtractor is an amateur product and is subject to further optimization. Future plans include:
- Development of a custom-made alignment model.
- Implementation of segmentation using U-Net models.
- Extensive performance optimizations.
Feedback and contributions to enhance and optimize this tool are greatly welcomed.