Diffusion GAN
Diffusion GANs combine the strengths of diffusion models and Generative Adversarial Networks (GANs) to improve image generation, aiming for high-fidelity, diverse outputs with faster inference speeds. Current research focuses on optimizing architectures for specific tasks like super-resolution, text-to-image synthesis, and 3D model generation, often incorporating techniques like wavelet transforms and latent space compression to enhance efficiency. These advancements are significant because they address limitations in both traditional GANs and diffusion models, leading to improved performance in various applications, including digital pathology, event-based vision, and music generation.
16papers
Papers
February 8, 2025
January 31, 2025
October 23, 2024
August 27, 2024
July 15, 2024
March 19, 2024
December 28, 2023
November 28, 2023
October 27, 2023
November 29, 2022