face_changing_v2

Final Project of CSCI 1430 with a self-implemented face landmark detection model

Install requirements
```
pip3 install -r requirements.txt
```
Dataset

(1) Wider Facial Landmarks in-the-wild (WFLW) is a new proposed face dataset. It contains 10000 faces (7500 for training and 2500 for testing) with 98 fully manual annotated landmarks.

(2) Download WFLW Training and Testing images at [Google Drive]

(3) Download WFLW Annotations at Face Annotations

(4) Unzip above two packages and put them on ./data/WFLW/

(5) Move Mirror98.txt to WFLW/WFLW_annotations

(6) Preprocess the dataset
```
$ cd data 
$ python3 SetPreparation.py
```

Train and Test

(1) Train:

$ python3 train.py

(2) Use tensorbnoard to view the loss changing:

$ tensorboard  --logdir=./checkpoint/tensorboard/

(3) Test:

$ python3 test.py

Face Changing

(1) Extract and save the video frames
```
$ python3 main.py
```
(2) Create the output video using saved frames
```
$ python3 transfer_to_video.py
```
Result
Report and Poster

Final report: https://drive.google.com/file/d/1hst6AY2cbunJz6_fLI4KINTOAoNz8cAf/view?usp=sharing

Poster: https://drive.google.com/file/d/1fDGP9QRNnN8bLNO9HHJ1RIW31amfG5Jz/view?usp=sharing
Reference

PFLD: A Practical Facial Landmark Detector https://arxiv.org/pdf/1902.10859.pdf

Name		Name	Last commit message	Last commit date
Latest commit History 20 Commits
checkpoint		checkpoint
conf		conf
data		data
dataset		dataset
face_change		face_change
images		images
masks		masks
models		models
mtcnn		mtcnn
output		output
pfld		pfld
.DS_Store		.DS_Store
README.md		README.md
face1.jpeg		face1.jpeg
face2.jpeg		face2.jpeg
face3.jpeg		face3.jpeg
face_putput.jpeg		face_putput.jpeg
indices		indices
main.py		main.py
requirements.txt		requirements.txt
result.py		result.py
show_mask.py		show_mask.py
test.py		test.py
train.py		train.py
transfer_to_video.py		transfer_to_video.py

HanCai98/Real-Time-Face-Swap