DLGAN: Disentangling Label-Specific Fine-Grained Features for Image Manipulation
arXiv:1911.09943
Abstract
Recent studies have shown how disentangling images into content and feature spaces can provide controllable image translation/ manipulation. In this paper, we propose a framework to enable utilizing discrete multi-labels to control which features to be disentangled, i.e., disentangling label-specific fine-grained features for image manipulation (dubbed DLGAN). By mapping the discrete label-specific attribute features into a continuous prior distribution, we leverage the advantages of both discrete labels and reference images to achieve image manipulation in a hybrid fashion. For example, given a face image dataset (e.g., CelebA) with multiple discrete fine-grained labels, we can learn to smoothly interpolate a face image between black hair and blond hair through reference images while immediately controlling the gender and age through discrete input labels. To the best of our knowledge, this is the first work that realizes such a hybrid manipulation within a single model. More importantly, it is the first work to achieve image interpolation between two different domains without requiring continuous labels as the supervision. Qualitative and quantitative experiments demonstrate the effectiveness of the proposed method.
References in corpus (14)
- Conditional Generative Adversarial Nets
- Large Scale GAN Training for High Fidelity Natural Image Synthesis
- Generative Adversarial Text to Image Synthesis
- Conditional Image Synthesis With Auxiliary Classifier GANs
- Toward Multimodal Image-to-Image Translation
- U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation
- Adversarial Autoencoders
- Disentangling factors of variation in deep representations using adversarial training
- Learning What and Where to Draw
- Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking
- Discovering Hidden Factors of Variation in Deep Networks
- Semantic Image Synthesis via Adversarial Learning
- Exemplar Guided Unsupervised Image-to-Image Translation with Semantic Consistency
- ELEGANT: Exchanging Latent Encodings with GAN for Transferring Multiple Face Attributes