Compact Latent Representation for Image Compression (CLRIC)
This addresses the issue of high training and storage costs for image compression models, making it more efficient for applications requiring multiple quality levels, though it is incremental as it builds on existing latent variable models.
The paper tackles the problem of resource-intensive image compression models that require separate models for each quality level by proposing a method that uses latent variables from pre-trained models like Stable Diffusion VAE, eliminating the need for distinct models and achieving comparable perceptual quality to state-of-the-art methods with low computational complexity of around 25.5 MAC/pixel.
Current image compression models often require separate models for each quality level, making them resource-intensive in terms of both training and storage. To address these limitations, we propose an innovative approach that utilizes latent variables from pre-existing trained models (such as the Stable Diffusion Variational Autoencoder) for perceptual image compression. Our method eliminates the need for distinct models dedicated to different quality levels. We employ overfitted learnable functions to compress the latent representation from the target model at any desired quality level. These overfitted functions operate in the latent space, ensuring low computational complexity, around $25.5$ MAC/pixel for a forward pass on images with dimensions $(1363 \times 2048)$ pixels. This approach efficiently utilizes resources during both training and decoding. Our method achieves comparable perceptual quality to state-of-the-art learned image compression models while being both model-agnostic and resolution-agnostic. This opens up new possibilities for the development of innovative image compression methods.