/QuantFace

QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization

Primary LanguagePython

This is the official repository of the paper:

QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization

Paper on arxiv: arxiv

Accepted ICPR 2022

frmework

Model Pretrained model
ResNet100 pretrained-mode
ResNet100 W8A8 (Real data) pretrained-mode
ResNet100 W8A8 (Synthetic data) pretrained-mode
ResNet100 W6A6 (Real data) pretrained-mode
ResNet100 W6A6 (Synthetic data) pretrained-mode
ResNet50 pretrained-mode
ResNet50 W8A8 (Real data) pretrained-mode
ResNet50 W8A8 (Synthetic data) pretrained-mode
ResNet50 W6A6 (Real data) pretrained-mode
ResNet50 W6A6 (Synthetic data) pretrained-mode
ResNet18 pretrained-mode
ResNet18 W8A8 (Real data) pretrained-mode
ResNet18 W8A8 (Synthetic data) pretrained-mode
ResNet18 W6A6 (Real data) pretrained-mode
ResNet18 W6A6 (Synthetic data) pretrained-mode
MobileFaceNet pretrained-mode
MobileFaceNet W8A8 (Real data) pretrained-mode
MobileFaceNet W8A8 (Synthetic data) pretrained-mode
MobileFaceNet W6A6 (Real data) pretrained-mode
MobileFaceNet W6A6 (Synthetic data) pretrained-mode

If you use any of the code provided in this repository, please cite the following paper:

Citation


@inproceedings{quantface_boutros,
	title        = {QuantFace: Towards Lightweight Face Recognition by Synthetic Data Low-bit Quantization},
	author       = {Fadi Boutros and Naser Damer and Arjan Kuijper},
	year         = 2022,
	booktitle = {26th International Conference on Pattern Recognition, {ICPR} 2022,
               Montreal,Quebec ,August 21-25, 2021},
        publisher = {{IEEE}},
        year      = {2022},
}

License

The dataset, the implementation, or trained models, use is restricted to research purpuses. The use of the dataset or the implementation/trained models for product development.
This project is licensed under the terms of the Attribution-NonCommercial-ShareAlike 4.0 
International (CC BY-NC-SA 4.0) license. 
Copyright (c) 2021 Fraunhofer Institute for Computer Graphics Research IGD Darmstadt