cs.LGNov 5, 2026

Flux as a Codec: Zero-Shot Lossy Compression with Flux

Authors: Fuma Kimishima, Jinjia Zhou

Organizations: Hosei University, Tokyo, Japan

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

We propose FluxC, a zero-shot codec that uses Flux as a compression prior. Although DiffC has shown that pretrained diffusion models can serve as zero-shot codecs, we ask how the choice of coding path and reconstruction method affects rate--distortion--perception behavior. We analyze FluxC along three axes: coding cost and bit allocation, reconstruction method, and low rate distortion. At matched signal-to-noise ratio (SNR), we show that the path choice does not affect the ideal information rate, but changes how bits are allocated across time steps. For reconstruction, we compare ancestral sampling, one-step ODE reconstruction, and many-step ODE reconstruction. Our analysis shows that ODE-based reconstruction can achieve lower distortion than ancestral sampling, especially in the low rate regime. Experiments on Gaussian sources and natural images support this prediction. Moreover, FluxC performs competitively with, and in some cases outperforms, recent neural image compression methods, suggesting that Flux can serve as a principled zero-shot codec.

Figures & tables

Figure 1
Figure 1: (Top) Rate as a function of the path time tt on three datasets. (Bottom) Rate as a function of SNR on a logarithmic scale.
Figure 2: Incremental rate as a function of tt . The incremental rate denotes the additional coding cost assigned to each coding step, i.e., the increase in rate between adjacent time points.
Figure 3: Signed MSE change relative to the corresponding zero rate reconstruction in the ultra low rate regime ( R≤10−3R\leq 10^{-3} bpp). More negative values indicate larger distortion reductions.
Figure 4: Rate–distortion and rate–perception curves on natural images. We report PSNR, LPIPS, and MS-SSIM as distortion and FID as perception on a logarithmic rate axis.
Figure 5: Qualitative reconstructions at approximately 0.010.01 bpp. The examples compare MS-ILLM ( Muckley et al., 2023 ) , DDCM ( Ohayon et al., 2025 ) , DiffC and FluxC variants
Appendix figures & tables8 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 6: Experiments on isotropic Gaussian source. Lower D/dD/d and lower P/dP/d are better.
Figure 7: Experiments on Gaussian source across mild power-law, heavy-tailed power-law, and low rank spectra.
DatasetMethodEnc stepsEncodeDecode
KodakDiffC SD1.55039.4172.82
KodakDiffC SD2.15038.2769.81
KodakDiffC Flux4636.00517.08
KodakFluxC-14627.35\mathbf{27.35}47.91\mathbf{47.91}
KodakFluxC-ODE4635.05574.51
DIV2KDiffC SD1.55040.1073.14
Appendix
Table 1: Encode and decode time on 512 ×\times 512 images across three datasets. Times are seconds per image. Encode and decode denote the average processing times per image. Bold indicates the lowest runtime for each dataset.
Figure 8: Estimation MSE of x0\mathbf{x}_{0} -prediction as a function of SNR for Flux, SD 1.5, and SD 2.1 on each dataset.
DatasetMethodEnc stepsEncodeDecode
KodakFluxC-ODE1611.7999.14
KodakFluxC-ODE64.5726.59
KodakFluxC-ODE45.099.65
KodakFluxC-11611.6718.48
KodakFluxC-164.45\mathbf{4.45}7.69
KodakFluxC-145.085.09\mathbf{5.09}
Appendix
Table 2: FluxC encoding and decoding time (s/image) on average. Bold indicates the lowest runtime for each dataset.
Figure 9: Rate–distortion and rate–perception curves for FluxC with nominal encoding schedules of 50, 20, 10, and 5 steps.
Figure 10: Additional qualitative reconstructions at approximately 0.0030.003 bpp compared with PerCo, DDCM, and DiffC variants and our FluxC variants.
Figure 11: Additional qualitative reconstructions at method-specific rates near the 0.020.02 bpp operating region.
More in this topic

Diffusion-Based Image Compression

See all 8 papers

Explore similar work

CardsList

  1. Gen2-IC: Bridging Generative Models and Image Codecs through Latent Transport

    Jun 19, 2026Yinhuan Huang, Hao Cao, Pu chen +3Learned Image CompressionGenerative Image Compression

    Tsinghua University

  2. Few-step Generative Models as Lossy Compression

    Jun 9, 2026Fuma Kimishima, Jinjia ZhouLearned Image CompressionFlow-Based Generative Modeling

    Hosei University, Tokyo, Japan

  3. BitFlat: Error-Localized Bitstream Design for Robust Diffusion Image Compression

    Oct 19, 2026Xinfeng Deng, Li Zhou, Haitao Chen +5Diffusion-Based Image CompressionGenerative Image Compression

    College of Electronic Science and Technology, National University of Defense Technology, Changsha, 410073, China.