< Antispoofing > Multi-Task-Learning in Face Antispoofing

Is Face Antispoofing only a binary classification task?

Of course, we can consider face antispoofing as a binary classification task. We can train a classifier to distinguish a face image between liveness and fake.

It may work well but it can not fully use all information of the input face image. In order to push the limit of our trained data and trained classifier, we have to cultivate other information that could help us to better discriminate an attack face.

So what information could cultivated from an image beyond its label?

The depth information of a face maybe helpful. The fake images which normally are paper, screen do not contain any depth information while the real face contains depth information. So we can use this extra depth information to separate the real face from fake face.

Multi-Task-Learning in Face antispoofing

Muti-Task-Learning means that a model could achieve its aim by learning other task simultaneously. Specially in deep learning, we could combine multiple loss together to train a deep model.

More specially in face antispoofing task, we could combine the depth loss and classification loss together to get a more generalized model.

To picture below shows the main process:

As for how to combine the classification loss and depth loss, you could try many ways(to add together or to optimize separately).
May this could help bring you some insight in your task.