Fidelity-Controllable Extreme Image Compression with Generative Adversarial Networks
This addresses image compression for applications requiring low bitrates, but it is incremental as it builds on existing GAN-based methods to improve stability and reduce artifacts.
The authors tackled the problem of blur in learned image compression at extremely low bitrates below 0.1bpp by proposing a GAN-based method with two-stage training and network interpolation, resulting in high-quality reconstructions preferred over state-of-the-art GAN-based models in a user study.
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 bitrates. Although GAN can help to reconstruct sharp images, there are two drawbacks. First, GAN makes training unstable. Second, the reconstructions often contain unpleasing noise or artifacts. To address both of the drawbacks, our method adopts two-stage training and network interpolation. The two-stage training is effective to stabilize the training. Moreover, the network interpolation utilizes the models in both stages and reduces undesirable noise and artifacts, while maintaining important edges. Hence, we can control the trade-off between perceptual quality and fidelity without re-training models. The experimental results show that our model can reconstruct high quality images. Furthermore, our user study confirms that our reconstructions are preferable to state-of-the-art GAN-based image compression model. The code will be available.