mlsanityData to 14 Oct 2026

Recasting DiT for Extreme Image Compression with Hierarchical Visual Priors

arXiv 2610.18607 ↗First posted 13 Oct 2026cs.CVcited by 0

On the map: Diffusion models → Diffusion for compression and wireless

Abstract (arXiv)

Pretrained diffusion models provide powerful generative priors for extreme image compression, but their stochastic, iterative generation processes are not designed to decode compressed representations directly. We present PrismCodec, which recasts a pretrained Diffusion Transformer (DiT) as a bitstream-grounded one-step generative decoder. PrismCodec uses the decoded representation in two complementary roles: a projected codec state defines the starting state for generation, while hierarchical visual priors derived from the same decoded latent provide semantic and structural guidance. To bridge the gap between stochastic, iterative generation and deterministic one-step decoding, we develop a progressive adaptation strategy. We first learn codec-conditioned transport from a noisy neighborhood centered on the codec state, distill it into a one-step mapping, and finally anneal the source perturbation to zero during end-to-end optimization. The resulting decoder reconstructs deterministically from the received representation without additional conditioning bits. Across Kodak, CLIC, and DIV2K, PrismCodec consistently improves rate-perception performance over prior generative codecs, achieving BD-rate reductions of 70-75% in LPIPS and 75-81% in DISTS over StableCodec. Code will be released upon acceptance.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Diffusion for compression and wireless, one of 14 directions in Diffusion models. Builds on 106 of its 124 references that are in the map. First posted on arXiv 13 Oct 2026.

Builds on

106 of 124 references.

Closest ancestors · 27

Same direction, newest first

  1. 2026CoD-Lite: Real-Time Diffusion-Based Generative Image Compression14 Apr
  2. 2026A Noise Constrained Diffusion (NC-Diffusion) Framework for High Fidelity Image Compression8 Apr
  3. 2026Next-Frame Decoding for Ultra-Low-Bitrate Image Compression with Video Diffusion Priors16 Mar
  4. 2026CADC: Content Adaptive Diffusion-Based Generative Image Compression25 Feb
  5. 2025CoD: A Diffusion Foundation Model for Image Compression24 Nov
  6. 2025Turbo-DDCM: Fast and Flexible Zero-Shot Diffusion-Based Image Compression9 Nov
  7. 2025Generative Image Coding with Diffusion Prior17 Sep
  8. 2025Steering One-Step Diffusion Model with Fidelity-Rich Decoder for Fast Image Compression7 Aug
  9. 2025StableCodec: Taming One-Step Diffusion for Extreme Image Compression27 Jun
  10. 2025One-Step Diffusion-Based Image Compression with Semantic Distillation22 May
  11. 2025OSCAR: One-Step Diffusion Codec Across Multiple Bit-rates22 May
  12. 2025Ultra Lowrate Image Compression with Semantic Residual Coding and Compression-aware Diffusion13 May
15 more
  1. 2025PICD: Versatile Perceptual Image Compression with Diffusion Rendering9 May
  2. 2025Bridging the Gap between Gaussian Diffusion Models and Universal Quantization for Image Compression3 Apr
  3. 2025Compressed Image Generation with Denoising Diffusion Codebook Models3 Feb
  4. 2025Lossy Compression with Pretrained Diffusion Models16 Jan
  5. 2024RDEIC: Accelerating Diffusion-Based Extreme Image Compression with Relay Residual Diffusion3 Oct
  6. 2024PerCo (SD): Open Perceptual Compression30 Sep
  7. 2024PSC: Posterior Sampling-Based Compression13 Jul
  8. 2024Towards Extreme Image Compression with Latent Feature Guidance and Diffusion Prior29 Apr
  9. 2024Lossy Image Compression with Foundation Diffusion Models12 Apr
  10. 2024Correcting Diffusion-Based Perceptual Image Compression with Privileged End-to-End Decoder7 Apr
  11. 2023High-Fidelity Image Compression with Score-based Generative Models26 May
  12. 2023A Residual Diffusion Model for High Perceptual Quality Codec Augmentation13 Jan
  13. 2022Extreme Generative Image Compression by Learning Text Embedding from Diffusion Models14 Nov
  14. 2022Lossy Image Compression with Conditional Diffusion Models14 Sep
  15. 2022Lossy Compression with Gaussian Diffusion17 Jun

Other references · 79

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  4. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  5. 2020Denoising Diffusion Implicit Modelscited by 9,781
  6. 2022Scalable Diffusion Models with Transformerscited by 6,054
  7. 2022Flow Matching for Generative Modelingcited by 5,702
  8. 2022Generative Adversarial Networkscited by 5,627
  9. 2023Adding Conditional Control to Text-to-Image Diffusion Modelscited by 5,503
  10. 2017Neural Discrete Representation Learningcited by 4,712
  11. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  12. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
38 more
  1. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  2. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,253
  3. 2025DINOv3cited by 1,414
  4. 2023T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Modelscited by 1,351
  5. 2018Demystifying MMD GANscited by 1,293
  6. 2023One-step Diffusion with Distribution Matching Distillationcited by 1,019
  7. 2018Variational image compression with a scale hyperpriorcited by 935
  8. 2020Image Quality Assessment: Unifying Structure and Texture Similaritycited by 811
  9. 2023Adversarial Diffusion Distillationcited by 759
  10. 2024Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Thinkcited by 703
  11. 2016End-to-end Optimized Image Compressioncited by 696
  12. 2024Improved Distribution Matching Distillation for Fast Image Synthesiscited by 684
  13. 2018Joint Autoregressive and Hierarchical Priors for Learned Image Compressioncited by 664
  14. 2020Learned Image Compression with Discretized Gaussian Mixture Likelihoods and Attention Modulescited by 590
  15. 2025Mean Flows for One-step Generative Modelingcited by 561
  16. 2025Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Modelscited by 393
  17. 2020High-Fidelity Generative Image Compressioncited by 338
  18. 2022ELIC: Efficient Learned Image Compression with Unevenly Grouped Space-Channel Contextual Adaptive Codingcited by 301
  19. 2020Channel-wise Autoregressive Entropy Models for Learned Image Compressioncited by 264
  20. 2018Generative Adversarial Networks for Extreme Learned Image Compressioncited by 219
  21. 2023Learned Image Compression with Mixed Transformer-CNN Architecturescited by 211
  22. 2021Checkerboard Context Model for Efficient Learned Image Compressioncited by 202
  23. 2022Entroformer: A Transformer-based Entropy Model for Learned Image Compressioncited by 107
  24. 2023Towards image compression with perfect realism at ultra-low bitratescited by 103
  25. 2023Improving Statistical Fidelity for Neural Image Compression with Implicit Local Likelihood Modelscited by 94
  26. 2025Latent Diffusion Model without Variational Autoencodercited by 88
  27. 2025Align Your Flow: Scaling Continuous-Time Flow Map Distillationcited by 87
  28. 2021Ensembling Off-the-shelf Models for GAN Trainingcited by 86
  29. 2022MLIC: Multi-Reference Entropy Model for Learned Image Compressioncited by 84
  30. 2022Multi-Realism Image Compression with a Conditional Generatorcited by 80
  31. 2023Text + Sketch: Image Compression at Ultra Low Ratescited by 67
  32. 2023MLIC++: Linear Complexity Multi-Reference Entropy Modeling for Learned Image Compressioncited by 58
  33. 2023Frequency-Aware Transformer for Learned Image Compressioncited by 58
  34. 2023EVC: Towards Real-Time Neural Image Compression with Mask Decaycited by 49
  35. 2024Posterior-Mean Rectified Flow: Towards Minimum MSE Photo-Realistic Image Restorationcited by 34
  36. 2022Image Compression with Product Quantized Masked Image Modelingcited by 33
  37. 2023ControlNet-XS: Rethinking the Control of Text-to-Image Diffusion Models as Feedback-Control Systemscited by 30
  38. 2025Learned Image Compression with Dictionary-based Entropy Modelcited by 30

50 most cited of 79 All 79 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026FlowCodec: One-Step Flow Prior for Generative Image Compression
  2. 2024Enhancing the Rate-Distortion-Perception Flexibility of Learned Image Codecs with Conditional Diffusion Decoders
  3. 2026Towards Efficient Low-rate Image Compression with Frequency-aware Diffusion Prior Refinement
  4. 2026Region-Adaptive Generative Compression with Spatially Varying Diffusion Models
  5. 2025Generative Preprocessing for Image Compression with Pre-trained Diffusion Models
  6. 2024Progressive Compression with Universally Quantized Diffusion Models
  7. 2025Toward Lightweight and Fast Decoders for Diffusion Models in Image and Video Generation
  8. 2026Rethinking Generative Image Compression at Extremely Low Bitrates