FaceX-Zoo

FaceX-Zoo copied to clipboard

你好，想请教一下，在进行AgeDB30数据集评估时，有一个04_FINAL_protocol_30_years.mat文件是什么样的格式？在哪里生成的呢？谢谢！

Yangna1

@wang21jun

BigcowPeking

SST训练时，每个epoch为什么要设置 exclude_id_set， 设置和不设置对性能影响大吗？ 如果训练样本的probe或gallery图像和queue里面图像是同一类，如何处理，代码里面有吗，谢谢

1

SST训练时，为什么是每个epoch设置 exclude_id_set， 设置和不设置对性能影响大吗？ 如果训练样本的probe或gallery图像和queue里面以前压入的图像是同一类，如何处理，代码里面有吗（不处理的话，就是当做负样本了吗），谢谢

vzhangmeng726

paper中写adopt SST scheme on the deep data (full version of MS1M-v1c)时， SST gains the leading accuracy in most of the test sets。 可以介绍更多训练deep data的训练细节吗？ 比如如何生成deep_data_shallow_train_list.txt。 每个ID 生成多少个正样本对（paper中说deep data 每个ID 40...

vzhangmeng726

Testsets for Masked face recognition?

1

zws98

Hi, I have a confusion while reading the code of test.py in DSDG project. As the paper mentioned, the final face anti-spoofing score is obtained by averaging the predicted values...

zhangxilun

最近尝试了几个公开的架构比如insightface/JDAI-FACEX/paddlePLSC+paddleClas，对比了一下，paddle提供了蒸馏案例，JDAI是否会考虑增加蒸馏相关案例呢？谢谢

CHANGErIBE

Augmentation for ferplus

1

Could you please share the agumentation you use for FERPlus dataset? Do you use the same as in AffectNet? https://github.com/JDAI-CV/FaceX-Zoo/blob/d3dac645a57028ccd76d88928a2949bdb9211c7a/addition_module/DMUE/pretrain/utils/make_transform.py#L7 Thank you!

toharl

Hello, While trying to convert the CDCN network to onnx, network gets messed up for some reason and cannot be used for inference later. Can I use the model convertor...

kadirbeytorun

Any reference about extract new texture from new mask?

9

Thks for ur excellent project, and the 3d face-mask is very robust. However, I want add more kinds of mask texture for face-mask generation but cannot found any documention about...

luameows