Synonymous Variational Inference for Perceptual Image Compression
This work addresses image compression for applications requiring perceptual quality, offering a theoretical framework that is incremental but integrates existing approaches.
The paper tackles perceptual image compression by proposing a synonymous variational inference method that re-analyzes the problem using semantic information theory, proving a triple tradeoff that covers existing schemes and demonstrating comparable performance with a single progressive codec.
Recent contributions of semantic information theory reveal the set-element relationship between semantic and syntactic information, represented as synonymous relationships. In this paper, we propose a synonymous variational inference (SVI) method based on this synonymity viewpoint to re-analyze the perceptual image compression problem. It takes perceptual similarity as a typical synonymous criterion to build an ideal synonymous set (Synset), and approximate the posterior of its latent synonymous representation with a parametric density by minimizing a partial semantic KL divergence. This analysis theoretically proves that the optimization direction of perception image compression follows a triple tradeoff that can cover the existing rate-distortion-perception schemes. Additionally, we introduce synonymous image compression (SIC), a new image compression scheme that corresponds to the analytical process of SVI, and implement a progressive SIC codec to fully leverage the model's capabilities. Experimental results demonstrate comparable rate-distortion-perception performance using a single progressive SIC codec, thus verifying the effectiveness of our proposed analysis method.