Self-Supervised CycleGAN for Object-Preserving Image-to-Image Domain Adaptation

Keyphrases

• Image Domain 100%
• Generative Adversarial Networks 100%
• CycleGAN 100%
• Domain Adaptation 100%
• Image-to-image 100%
• Image-to-image Translation 40%
• Segmentation Network 20%
• Image Content 20%
• Image Object 20%
• Segmentation Accuracy 20%
• Content Consistency 20%
• Publicly Available Dataset 20%
• Deep Learning Network 20%
• Network-based Methods 20%
• Semantic Segmentation 20%
• Pixel-wise 20%
• PSPNet 20%
• Adaptation Scenarios 20%

Computer Science

• Generative Adversarial Networks 100%
• Domain Adaptation 100%
• Annotation 40%
• Image Translation 40%
• Deep Learning 20%
• Experimental Result 20%
• Image Segmentation 20%
• Image Content 20%
• segmentation accuracy 20%
• Regularization 20%
• Network Segmentation 20%