Generative Face Completion
This work addresses face completion for applications in image editing and restoration, but it is incremental as it builds on existing deep generative methods.
The paper tackles the problem of face completion, which involves generating semantically new pixels for missing key components like eyes and mouths, and demonstrates that their model can handle large missing areas in arbitrary shapes to produce realistic results.
In this paper, we propose an effective face completion algorithm using a deep generative model. Different from well-studied background completion, the face completion task is more challenging as it often requires to generate semantically new pixels for the missing key components (e.g., eyes and mouths) that contain large appearance variations. Unlike existing nonparametric algorithms that search for patches to synthesize, our algorithm directly generates contents for missing regions based on a neural network. The model is trained with a combination of a reconstruction loss, two adversarial losses and a semantic parsing loss, which ensures pixel faithfulness and local-global contents consistency. With extensive experimental results, we demonstrate qualitatively and quantitatively that our model is able to deal with a large area of missing pixels in arbitrary shapes and generate realistic face completion results.