Generative Face Completion

Abstract

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.

Accepted by CVPR 2017

References in corpus (1)

• Very Deep Convolutional Networks for Large-Scale Image Recognition

Cited by in corpus (34)

• Adversarial Learning for Semi-Supervised Semantic Segmentation
• Generative Image Inpainting with Contextual Attention
• Deep Learning in Mobile and Wireless Networking: A Survey
• SPG-Net: Segmentation Prediction and Guidance Network for Image Inpainting
• Attentive Generative Adversarial Network for Raindrop Removal from a Single Image
• Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks
• Shift-Net: Image Inpainting via Deep Feature Rearrangement
• SC-FEGAN: Face Editing Generative Adversarial Network with User's Sketch and Color
• FSRNet: End-to-End Learning Face Super-Resolution with Facial Priors
• Structural inpainting
• SemanticAdv: Generating Adversarial Examples via Attribute-conditional Image Editing
• A Pixel-Based Framework for Data-Driven Clothing
• Identity Preserving Face Completion for Large Ocular Region Occlusion
• Discriminative Region Proposal Adversarial Networks for High-Quality Image-to-Image Translation
• Recent Progress of Face Image Synthesis
• Semantically Consistent Image Completion with Fine-grained Details
• Face Completion with Semantic Knowledge and Collaborative Adversarial Learning
• Learning Disentangling and Fusing Networks for Face Completion Under Structured Occlusions
• What Is It Like Down There? Generating Dense Ground-Level Views and Image Features From Overhead Imagery Using Conditional Generative Adversarial Networks
• SD-GAN: Structural and Denoising GAN reveals facial parts under occlusion
• Geometry-Aware Face Completion and Editing
• Image Quality Assessment Techniques Show Improved Training and Evaluation of Autoencoder Generative Adversarial Networks
• Data recovery in computational fluid dynamics through deep image priors
• Play as You Like: Timbre-enhanced Multi-modal Music Style Transfer
• Multiple Sclerosis Lesion Inpainting Using Non-Local Partial Convolutions
• Webpage Saliency Prediction with Two-stage Generative Adversarial Networks
• High-dimensional Assisted Generative Model for Color Image Restoration
• Learning Symmetry Consistent Deep CNNs for Face Completion
• Convolutional Point-set Representation: A Convolutional Bridge Between a Densely Annotated Image and 3D Face Alignment
• One-Stage Inpainting with Bilateral Attention and Pyramid Filling Block
• Escaping from Collapsing Modes in a Constrained Space
• Generating a Fusion Image: One's Identity and Another's Shape
• Cascade context encoder for improved inpainting
• Association: Remind Your GAN not to Forget