MaCo-GAN: Manifold-Contrastive Adversarial Learning for Single Image Super-Resolution
Researchers have introduced MaCo-GAN, a new framework for single image super-resolution that utilizes manifold-contrastive adversarial learning. This approach replaces traditional adversarial loss with a supervised contrastive objective, generating a spectrum of realistic fake images that maintain low-resolution correspondence. The generator is trained to align its predictions with on-manifold fakes while repelling off-manifold ones, leading to improved perception-distortion trade-offs. AI
IMPACT Introduces a novel contrastive learning approach that could improve image quality in super-resolution tasks.