Enjoy Your Editing: Controllable GANs for Image Editing via Latent Space Navigation
arXiv:2102.01187
Abstract
Controllable semantic image editing enables a user to change entire image attributes with a few clicks, e.g., gradually making a summer scene look like it was taken in winter. Classic approaches for this task use a Generative Adversarial Net (GAN) to learn a latent space and suitable latent-space transformations. However, current approaches often suffer from attribute edits that are entangled, global image identity changes, and diminished photo-realism. To address these concerns, we learn multiple attribute transformations simultaneously, integrate attribute regression into the training of transformation functions, and apply a content loss and an adversarial loss that encourages the maintenance of image identity and photo-realism. We propose quantitative evaluation strategies for measuring controllable editing performance, unlike prior work, which primarily focuses on qualitative evaluation. Our model permits better control for both single- and multiple-attribute editing while preserving image identity and realism during transformation. We provide empirical results for both natural and synthetic images, highlighting that our model achieves state-of-the-art performance for targeted image manipulation.
arXiv commentAccepted to ICLR 2021. 14 pages, 15 figures
References in corpus (13)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Self-Attention Generative Adversarial Networks
- Progressive Growing of GANs for Improved Quality, Stability, and Variation
- The relativistic discriminator: a key element missing from standard GAN
- GANSpace: Discovering Interpretable GAN Controls
- Semantic Image Synthesis with Spatially-Adaptive Normalization
- Unsupervised Discovery of Interpretable Directions in the GAN Latent Space
- Controllable Text-to-Image Generation
- Interpreting the Latent Space of GANs for Semantic Face Editing
- Closed-Form Factorization of Latent Semantics in GANs
- On the "steerability" of generative adversarial networks
- StyleGAN2 Distillation for Feed-forward Image Manipulation