Facial Landmark Detection Using Generative Adversarial Network Combined with Autoencoder for Occlusion

Figure 3

The comparison of facial occluded image recovery with or without self-attention mechanism. (a) Original image, (b) the occluded image, (c) the left eye and right eye are asymmetric of the third image recovered face without self-attention mechanism, and (d) recovered face without self-attention mechanism.

(a)

(b)

(c)

(d)