17 citations · 17 across the 2 of their papers we have counts for
3 papers
Layout-Corrector: Alleviating Layout Sticking Phenomenon in Discrete Diffusion Model
Shoma Iwai, Atsuki Osanai, Shunsuke Kitada +1
Layout generation is a task to synthesize a harmonious layout with elements characterized by attributes such as category, position, and size. Human designers experiment with the pl…
Controlling Rate, Distortion, and Realism: Towards a Single Comprehensive Neural Image Compression Model
Shoma Iwai, Tomo Miyazaki, Shinichiro Omachi
In recent years, neural network-driven image compression (NIC) has gained significant attention. Some works adopt deep generative models such as GANs and diffusion models to enhanc…
Fidelity-Controllable Extreme Image Compression with Generative Adversarial Networks
Shoma Iwai, Tomo Miyazaki, Yoshihiro Sugaya +1
We propose a GAN-based image compression method working at extremely low bitrates below 0.1bpp. Most existing learned image compression methods suffer from blur at extremely low bi…