https://github.com/Sanjana-Sarda/FVB

Face Verification Bypass
https://github.com/Sanjana-Sarda/FVB

Last synced: about 1 month ago
JSON representation

Face Verification Bypass

• Host: GitHub
• URL: https://github.com/Sanjana-Sarda/FVB
• Owner: Sanjana-Sarda
• Created: 2022-02-17T02:37:09.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2022-03-10T08:36:09.000Z (almost 3 years ago)
• Last Synced: 2024-08-12T08:09:28.332Z (5 months ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 34.6 MB
• Stars: 10
• Watchers: 3
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-MLSecOps - FVB

README

        
# Face Verification Bypass

Evaluate data on baseline using Generate_Profile_Images_baseline.

Select appropriate results and evaluate on Face_Verification.

## About The Project

Generating fake profile images that can be verified with your own face but look as different as possible.

## Getting Started

All files can be run locally using jupyter.

## Usage

1. Preprocess images from your personal dataset by running Data_Preprocessing.ipynb. 

2. Set up the face verification system by running Face_Verification.ipynb. Save the generated database.pkl file for later use.

3. To test baseline use Generate_Profile_Images_baseline.ipynb. This is based on FreezeD. 

4. To test on starGAN v2 use starGAN_v2.ipynb. 

5. Evaluate selected generated images in Face_Verification.ipynb under Experiments by loading the previously downloaded database.pkl file. 

6. To experiment with a naive Freeze-D implementation for starGAN, replace the solver.py file in the core folder (after downloading the repository) with the solver.py file in this repository.