Need help with FaceQnet?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

uam-biometrics
187 Stars 49 Forks 28 Commits 2 Opened issues

Description

FaceQnet: Quality Assessment for Face Recognition based on Deep Learning

Services available

!
?

Need anything else?

Contributors list

No Data

FaceQnet

FaceQnet: Quality Assessment for Face Recognition based on Deep Learning

This repository currently contains two different versions of FaceQnet:

FaceQnet is a No-Reference, end-to-end Quality Assessment (QA) system for face recognition based on deep learning. The system consists of a Convolutional Neural Network that is able to predict the suitability of a specific input image for face recognition purposes. The training of FaceQnet is done using the VGGFace2 database.

-- Configuring environment in Windows:

1) Installing Conda: https://conda.io/projects/conda/en/latest/user-guide/install/windows.html

Update Conda in the default environment:

conda update conda
conda upgrade --all

Create a new environment:

conda create -n [env-name]

Activate the environment:

conda activate [env-name]

2) Installing dependencies in your environment:

Install Tensorflow and all its dependencies:

pip install tensorflow

Install Keras:

pip install keras

Install OpenCV:

conda install -c conda-forge opencv

3) If you want to use a CUDA compatible GPU for faster predictions:

You will need CUDA and the Nvidia drivers installed in your computer: https://docs.nvidia.com/deeplearning/sdk/cudnn-install/

Then, install the GPU version of Tensorflow:

pip install tensorflow-gpu

-- Using FaceQnet for predicting scores:

1) Download or clone the repository. 2) Due to the size of the video example, please download one of the the FaceQnet pretrained models and place the downloaded .h5 file it in the /src folder:

3) Edit and run the FaceQNetobtainscoresKeras.py script. - You will need to change the folder from which the script will try to charge the face images. It is src/Samplescropped by default. - The best results will be obtained when the input images have been cropped just to the zone of the detected face. In our experiments we have used the MTCNN face detector from <a href="https://kpzhang93.github.io/MTCNNfacedetectionalignment/index.html" rel="nofollow">here, but other detector can be used. - FaceQnet will ouput a quality score for each input image. All the scores will are saved in a .txt file into the src folder. This file contain each filename with its associated quality metric.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.