Frequency-Aware Perceptual Optimization for Low-Complexity Implicit Image Compression

We propose a frequency-aware perceptual optimization framework for low-complexity image compression, realized as a Realism-enhanced Region-based Implicit Codec (Re2IC). Re2IC models visual perception via saliency-guided region partitioning and local–global perceptual modulation. To enhance realism under complexity constraints, we introduce wavelet–Wasserstein distortion (WA-WD), a frequency-decomposed perceptual distortion that balances fidelity and realism through subband-aware modeling and provides a more reliable approximation than standard Wasserstein distortion. Together, these designs enable fine-grained spatial–spectral optimization, allowing Re2IC to achieve superior rate–perception trade-offs, outperforming generative codecs such as HiFiC while using less than 1\% of their decoding cost. Extensive experiments demonstrate state-of-the-art perceptual performance among overfitted codecs. Beyond compression, WA-WD serves as a standalone, tunable perceptual metric with strong alignment to human preference (Pearson 94.6\%, Spearman 92.3\%) and competitive performance across multiple IQA benchmarks. Project page: https://eedavidwu.github.io/ReReIC/

Image compression saves storage and transmission costs, but compressed images can look blurry or unnatural. We introduce Re$^2$IC, a lightweight image compression method that aims to preserve realistic visual quality with very low decoding cost. Re$^2$IC is inspired by how people perceive images: important regions, edges, textures, and smooth areas should be treated differently. To support this, we introduce WA-WD, a new image-quality measure that evaluates visual differences across both image regions and frequency patterns. This helps balance faithfulness to the original image with a realistic appearance. Experiments show that Re$^2$IC achieves strong perceptual quality and can outperform generative compression methods while using less than $1\%$ of their decoding cost. WA-WD also works as a general image-quality measure and aligns well with human preferences.