semi-adversarial-networks - GitLite

KO | EN

gitlite — search

#python #llm #go #rust #deep-learning #java #windows #awesome #cli #kubernetes #vue #pytorch #security #open-source

semi-adversarial-networks

(★ 63)

Implementation of the SAN model for imparting gender privacy to face images

파일 탐색기

최종 버전 다운로드 (.zip)

- evaluate-SAN-model.ipynb
- aux-gpred.pkl
- conv-autoencoder-e20.pkl
- README.md
- vgg_gray_corr_ptch.npz
- convae.py
- dataset_loader.py
- genderpred.py
- main.py
- vggface.py
.gitignore
LICENSE
README.md
requirements.txt

# 설치 가이드

1. 코드 내려받기

git clone https://github.com/iPRoBe-lab/semi-adversarial-networks

깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.

cd semi-adversarial-networks

방금 내려받은 프로젝트 폴더 안으로 이동합니다.

2. Python

쉬움 추천

사전 준비물

Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
Python 3 설치 시 'Add Python to PATH' 옵션을 꼭 체크하세요 (Windows).

pip install -r requirements.txt

requirements.txt 등에 명시된 파이썬 라이브러리를 설치합니다.

jupyter notebook

브라우저에서 노트북(.ipynb) 파일들을 열람하고 실행할 수 있는 Jupyter 화면을 켭니다.

✅ 에러 메시지 없이 실행되고 터미널에 안내 문구가 출력되면 정상입니다.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

Semi Adversarial Networks (SAN)
===============================

PyTorch implementation of the semi-adversarial neural network model described in *Semi-adversarial networks: Convolutional autoencoders for imparting privacy to face images* by V. Mirjalili, S. Raschka, A. Namboodiri, and A. Ross.

## Research Article

[Vahid Mirjalili](https://github.com/vmirly), [Sebastian Raschka](https://github.com/rasbt), [Anoop Namboodiri](https://www.iiit.ac.in/people/faculty/anoop/), and [Arun Ross](http://www.cse.msu.edu/~rossarun/) (2018) *Semi-adversarial networks: Convolutional autoencoders for imparting privacy to face images.* Proc. of 11th IAPR International Conference on Biometrics (ICB 2018), Gold Coast, Australia.

- Arxiv preprint: [https://arxiv.org/abs/1712.00321](https://arxiv.org/abs/1712.00321)

## Implementation details and requirements

The model was implemented in PyTorch 0.3.1 using Python 3.6 and may be compatible with different versions of PyTorch and Python, but it has not been tested.

Additional requirements are listed in the [./requirements.txt](./requirements.txt) file.

## Usage

**Source code and model parameters**

The source code of the SAN model can be found in the [src](./src) subdirectory, and the parameters of the trained models are available in the [model](./model) subdirectory.

**Dataset**

The original dataset was obtained from http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html, but the images in the dataset were cropped using DPM model (source code can be found in http://www.robots.ox.ac.uk/~vgg/software/vgg_face/). In addition, we have aligned the prototypes to each input image.

**Training a SAN model**

In order to train a SAN model as described in the research paper, 
download the preprocessed image files (see previous paragraph) and place them into the main directory of this GitHub repository. To initiate the model training, execute the following code in your shell terminal.

```bash
cd src/
python main.py --gpu
```

Note that the `--gpu` flag specifies which GPU on your system shall be used; `--gpu 0` uses the first CUDA GPU device, `--gpu 1` uses the second CUDA GPU device and so forth.

**Examples**

Please see the [examples/evaluate-SAN-model.ipynb](examples/evaluate-SAN-model.ipynb) Jupyter Notebook for an example that illustrates how to instantiate the trained autoencoder part of the SAN model for evaluation, using the available model parameters.

More usage examples may follow in future.