MobileLiteNets for Face Anti-spoofing Attack Detection Challenge@CVPR2019[1]
TODO: Describe what is MobileLiteNets. How we explain FeatherNet?
We use 2 dataset to train. One is CASIA-SURF Dataset, another is created by ourselves, named Multi-Modality Face Dataset', abbreviated as MMFD.
TODO: data preparation, training script, where is the training model located
Commands to train the model:
python main.py --config="cfgs/MobileLiteNet54-32.yaml" --every-decay 60 -b 32 --lr 0.01 --fl-gamma 3 >>FNet54-bs32-train.log
python main.py --config="cfgs/MobileLiteNet54-se-64.yaml" --b 64 --lr 0.01 --every-decay 60 --fl-gamma 3 >> FNet54-se-bs64-train.log
python main.py --config="cfgs/FeatherNetA-32.yaml" --b 32 --lr 0.01 --every-decay 60 --fl-gamma 3 >> MobileLiteNetA-bs32-train.log
python main.py --config="cfgs/FeatherNetB-32.yaml" --b 32 --lr 0.01 --every-decay 60 --fl-gamma 3 >> MobileLiteNetB-bs32--train.log
TODO: Describe where is the trained model located?
TODO: framework used, model size, performance data (can use the following table as performance data)
model name | ACER | TPR@FPR=10E-2 | TPR@FPR=10E-3 | FP | FN | epoch | params | FLOPs |
---|---|---|---|---|---|---|---|---|
MobileLiteNet54 | 0.00242 | 1.0 | 0.99846 | 32 | 0 | 41 | 0.57M | 270.91M |
MobileLiteNet54-se | 0.00242 | 1.0 | 0.996994 | 32 | 0 | 69 | 0.57M | 270.91M |
FeatherNetA | 0.00261 | 1.00 | 0.961590 | 19 | 7 | 51 | 0.35M | 79.99M |
FeatherNetB | 0.00168 | 1.0 | 0.997662 | 20 | 1 | 48 | 0.35M | 83.05M |
TODO: Inference time for FeatherNet and MobileLiteNet
[1] ChaLearn Face Anti-spoofing Attack Detection Challenge@CVPR2019,link [2] Shifeng Zhang, Xiaobo Wang, Ajian Liu, Chenxu Zhao, Jun Wan, Sergio Escalera, Hailin Shi, Zezheng Wang, Stan Z. Li, " CASIA-SURF: A Dataset and Benchmark for Large-scale Multi-modal Face Anti-spoofing ", arXiv, 2018 PDF