KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #dsh #dsh-plugin #ai-agent #claude-code #codex #typescript #ai-agents #mcp #cli #rust #deepseek-harness

Talking-Face-Generation-DAVS

★ 813 Open GitHub ↗

Code for Talking Face Generation by Adversarially Disentangled Audio-Visual Representation (AAAI 2019)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Talking Face Generation by Adversarially Disentangled Audio-Visual Representation (AAAI 2019)

We propose `Disentangled Audio-Visual System (DAVS)` to address arbitrary-subject talking face generation in this work, which aims to synthesize a sequence of face images
that correspond to given speech semantics, conditioning on either an unconstrained speech audio or video.

[[Project]](https://liuziwei7.github.io/projects/TalkingFace) [[Paper]](https://arxiv.org/abs/1807.07860) [[Demo]](https://www.youtube.com/watch?v=-J2zANwdjcQ)

### Recommondation of our CVPR21 repo
This repo is barely maintaining since the version of this code is out of date. If you are interested in the topic of Talking Face Generation, feel free to try the [CODE](https://github.com/Hangz-nju-cuhk/Talking-Face_PC-AVS) of our [CVPR2021 PAPER](https://hangz-nju-cuhk.github.io/projects/PC-AVS)!

## Requirements
* [python 2.7](https://www.python.org/download/releases/2.7/)
* [PyTorch](https://pytorch.org/)（We use version 0.2.0)
* [opencv2](https://opencv.org/releases.html)

## Generating test results
* Download the pre-trained model [checkpoint](https://drive.google.com/open?id=1UW22xm4r9AewNoySyPd2fyUab0nqymBR)
``` bash
Create the default folder "checkpoints" and put the checkpoint in it or get the CHECKPOINT_PATH
```

* Samples for testing can be found in this folder named [0572_0019_0003](https://drive.google.com/open?id=1ykjOZwwFfyP2V1vdUVsm2v4r1QSM-uxa). This is a pre-processed sample from the [Voxceleb](http://www.robots.ox.ac.uk/~vgg/data/voxceleb/) Dataset.

* Run the testing script to generate videos from video:

``` bash
python test_all.py  --test_root ./0572_0019_0003/video --test_type video --test_audio_video_length 99 --test_resume_path CHECKPOINT_PATH
```
* Run the testing script to generate videos from audio:
``` bash
python test_all.py  --test_root ./0572_0019_0003/audio --test_type audio --test_audio_video_length 99 --test_resume_path CHECKPOINT_PATH
```

## Sample Results
* Talking Effect on Human Characters
<img src='./misc/demo_human.gif' width=640>

* Talking Effect on Non-human Characters (Trained on Human Faces Only)
<img src='./misc/demo_nonhuman.gif' width=640>

## Create more samples

* The face detection tool used in the demo videos can be found at [RSA](https://github.com/sciencefans/RSA-for-object-detection). It will return a Matfile with 5 key point locations in a row for each image. Other face alignment methods are also appliable such as [dlib](http://blog.dlib.net/2014/08/real-time-face-pose-estimation.html). The key points for face alignement we used are the two for the center of the eyes and the average point of the corners of the mouth. With each image's PATH and the face POINTS, you can find our way of face alignment at `preprocess/face_align.py`.

* Our preprocessing of the audio files is the same and borrowed from the matlab code of [SyncNet](http://www.robots.ox.ac.uk/~vgg/software/lipsync/). Then we save the mfcc features into bin files.

## Preparing Training Data
* We used the [LRW](http://www.robots.ox.ac.uk/~vgg/data/lip_reading/lrw1.html) dataset for training.
* The directories are arranged like this:
```
data
├── train, val, test
|	├── 0, 1, 2 ... 499 (one folder for each class)
|	│   ├── 0, 1, 2 ... #videos per class
|	│   │   ├── align_face256
|	│   │   |   ├── 0, 1, ... 28.jpg
|	│   |   ├── mfcc20
|	│   │   |   ├── 2, 3 ... 26.bin
```
where each video is extracted to frames and aligned using our protocol, and each audio is processed and saved using Matlab.

## Training 
```bash
python train.py
```
* This is still a beta version of the training code which only disentangles wid information from pid space. Running the train.py only might not be able to fully reproduce the paper. However, it can be served as a reference for how we implement the whole training process.
* During our own implementation, the classification part (without generation and disentanglement) is pretrained first. The pretraining training code is temporarily not provided.

## Postprocessing Details (Optional)

* The directly generated results may suffer from a "zoom-in-and-out" condition which we assume is caused by our alignment of the training set. We solve the unstable problem using [Subspace Video Stabilization](http://web.cecs.pdx.edu/~fliu/papers/tog2010.pdf) in the demos.

## License and Citation
The use of this software is RESTRICTED to **non-commercial research and educational purposes**.

```
@inproceedings{zhou2019talking,
  title     = {Talking Face Generation by Adversarially Disentangled Audio-Visual Representation},
  author    = {Zhou, Hang and Liu, Yu and Liu, Ziwei and Luo, Ping and Wang, Xiaogang},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year      = {2019},
}
```

## Acknowledgement
The structure of this codebase is borrowed from [pix2pix](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix).

Talking-Face-Generation-DAVS

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README