Implementation of DUL (PyTorch version)

Introduction

---

This repo is an unofficial PyTorch implementation of DUL (Data Uncertainty Learning in Face Recognition, CVPR2020).

NOTE:

1. SE-Resnet64 is used as defult backbone in this repo, you can define others in ./backbone/model_irse.py
2. Training (process) & Testing (results) logs are saved in ./exp/logs/ & ./exp/logs_test/
3. Implementation details are not exactly the same as the original paper, seen in ./config.py

Getting Started

---

• Star this repo, plz

  😊

• Clone this repo

git clone https://github.com/MouxiaoHuang/DUL.git

• Prepare env

conda create --name <env_name> python=3.8
pip install -r requirements.txt

• Prepare trainset and testset
  • Trainset: Casia WebFace or MS-Celeb-1M
  • Testset: LFW, CFP_FF, CFP_FP, AgeDB, CALFW, CPLFW, VGG2_FP
• Training

sh ./exp/Exp_webface_DUL.sh
# or
sh ./exp/Exp_ms1m_DUL.sh

• Testing

sh ./exp/TestFR_webface_DUL.sh
# or
sh ./exp/TestFR_ms1m_DUL.sh

Results Report

---

• Trainset: Casia Webface

	LFW	CFP_FF	CFP_FP	AgeDB	CALFW	CPLFW	VGG2_FP
Original paper	-	-	-	-	-	-	-
This repo	99.42	99.23	96.53	93.93	93.48	89.60	93.76

• Trainset: MS-Celeb-1M

	LFW	CFP_FF	CFP_FP	AgeDB	CALFW	CPLFW	VGG2_FP
Original paper (ResNet64)	99.78	-	98.67	-	-	-	-
This repo	99.75	99.69	98.41	98.02	95.95	92.97	95.40

Thanks & Refs

---

• ZhaoJ9014/face.evoLVe
• Ontheway361/dul-pytorch