标题： 同义变分推理用于感知图像压缩 显示英文标题

标题： Synonymous Variational Inference for Perceptual Image Compression

摘要： 语义信息论的最新贡献揭示了语义信息与句法信息之间的集合-元素关系，这种关系表现为同义关系。 本文基于这种同义性视角提出了一种同义变分推理（SVI）方法，重新分析感知图像压缩问题。 它以感知相似性作为典型的同义标准，构建理想的同义集（Synset），并通过最小化部分语义KL散度，用参数密度近似其潜在同义表示的后验分布。 该分析从理论上证明，感知图像压缩的优化方向遵循一种可以涵盖现有率失真感知方案的三元权衡。 此外，我们引入了同义图像压缩（SIC），这是一种对应于SVI分析过程的新图像压缩方案，并实现了一个渐进式SIC编解码器，以充分利用模型的能力。 实验结果表明，使用单一渐进式SIC编解码器可获得相当的率失真感知性能，从而验证了我们所提出的分析方法的有效性。 显示英文摘要

摘要： 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.