Search papers, labs, and topics across Lattice.
This paper introduces a hierarchical latent representation for learned image compression that utilizes multiple latents at different spatial scales, each equipped with its own entropy model. This method addresses the limitations of traditional single-latent systems by more effectively capturing the spatial structure of images. Experimental results indicate a significant 17.9% BD-rate reduction over VVC on the Kodak dataset, showcasing the potential of multi-scale latent representations in enhancing compression efficiency.
Achieving a 17.9% reduction in BD-rate over VVC demonstrates that multi-scale latent representations can significantly enhance image compression efficiency.
Most learned image compression systems rely on a single latent representation combined with a hyperprior, which limits their ability to efficiently capture image structure across spatial scales. In this work, we propose a hierarchical latent representation to improve the efficiency of the entropy model. By using multiple latents at different scales, each with its own entropy model, we better capture the spatial structure of the latent representation. Our experiments show that this approach achieves a 17.9% BD-rate reduction over VVC on Kodak, demonstrating the effectiveness of multi-scale latent representations. Furthermore, the approach is orthogonal to other advances in learned image compression, making it a versatile addition to existing methods.
