mlsanityData to 7 Oct 2026

Rethinking Lossy Compression: The Rate-Distortion-Perception Tradeoff

arXiv 1901.07821 ↗First posted 23 Jan 2019cs.LG cs.CV cs.ITcited by 207

Not on the map. Why?

Cited by 52 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Lossy compression algorithms are typically designed and analyzed through the lens of Shannon's rate-distortion theory, where the goal is to achieve the lowest possible distortion (e.g., low MSE or high SSIM) at any given bit rate. However, in recent years, it has become increasingly accepted that "low distortion" is not a synonym for "high perceptual quality", and in fact optimization of one often comes at the expense of the other. In light of this understanding, it is natural to seek for a generalization of rate-distortion theory which takes perceptual quality into account. In this paper, we adopt the mathematical definition of perceptual quality recently proposed by Blau & Michaeli (2018), and use it to study the three-way tradeoff between rate, distortion, and perception. We show that restricting the perceptual quality to be high, generally leads to an elevation of the rate-distortion curve, thus necessitating a sacrifice in either rate or distortion. We prove several fundamental properties of this triple-tradeoff, calculate it in closed form for a Bernoulli source, and illustrate it visually on a toy MNIST example.

Key facts. Cited by 207 papers in the map (data through 7 Oct 2026). First posted on arXiv 23 Jan 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 17; 2022: 19; 2023: 18; 2024: 51; 2025: 49; 2026: 53.

207 papers.

Sort:

By area:

  1. 2022ELIC: Efficient Learned Image Compression with Unevenly Grouped Space-Channel Contextual Adaptive Codingcited by 296
  2. 2021High-Fidelity GAN Inversion for Image Attribute Editingcited by 197
  3. 2022Beyond Transmitting Bits: Context, Semantics, and Task-Oriented Communicationscited by 166
  4. 2022Lossy Image Compression with Conditional Diffusion Modelscited by 165
  5. 2022The Devil Is in the Details: Window-based Attention for Image Compressioncited by 141
  6. 2023Towards image compression with perfect realism at ultra-low bitratescited by 99
  7. 2023Improving Statistical Fidelity for Neural Image Compression with Implicit Local Likelihood Modelscited by 90
  8. 2022Lossy Compression with Gaussian Diffusioncited by 87
  9. 2022Multi-Realism Image Compression with a Conditional Generatorcited by 79
  10. 2023Text + Sketch: Image Compression at Ultra Low Ratescited by 63
  11. 2024Towards Extreme Image Compression with Latent Feature Guidance and Diffusion Priorcited by 56
  12. 2022Generative Joint Source-Channel Coding for Semantic Image Transmissioncited by 55
38 more
  1. 2023NOFA: NeRF-based One-shot Facial Avatar Reconstructioncited by 53
  2. 2021A Theory of the Distortion-Perception Tradeoff in Wasserstein Spacecited by 52
  3. 2021Lossy Compression for Lossless Predictioncited by 51
  4. 2021Universal Rate-Distortion-Perception Representations for Lossy Compressioncited by 46
  5. 2022No Free Lunch Theorem for Security and Utility in Federated Learningcited by 39
  6. 2021Neural Video Compression using GANs for Detail Synthesis and Propagationcited by 35
  7. 2025Flow to the Mode: Mode-Seeking Diffusion Autoencoders for State-of-the-Art Image Tokenizationcited by 35
  8. 2024Lossy Image Compression with Foundation Diffusion Modelscited by 34
  9. 2024Autoregressive Diffusion Transformer for Text-to-Speech Synthesiscited by 34
  10. 2023A Residual Diffusion Model for High Perceptual Quality Codec Augmentationcited by 33
  11. 2025Token Communications: A Large Model-Driven Framework for Cross-modal Context-aware Semantic Communicationscited by 32
  12. 2021A coding theorem for the rate-distortion-perception functioncited by 30
  13. 2024MISC: Ultra-low Bitrate Image Semantic Compression Driven by Large Multimodal Modelcited by 29
  14. 2025StableCodec: Taming One-Step Diffusion for Extreme Image Compressioncited by 29
  15. 2021On the advantages of stochastic encoderscited by 28
  16. 2021On Perceptual Lossy Compression: The Cost of Perceptual Reconstruction and An Optimal Training Frameworkcited by 27
  17. 2024Idempotence and Perceptual Image Compressioncited by 27
  18. 2024Latency-Aware Generative Semantic Communications with Pre-Trained Diffusion Modelscited by 25
  19. 2021Optimal Transport for Unsupervised Denoising Learningcited by 24
  20. 2025Compressed Image Generation with Denoising Diffusion Codebook Modelscited by 24
  21. 2025One-Step Diffusion-Based Image Compression with Semantic Distillationcited by 23
  22. 2024Interpretability as Compression: Reconsidering SAE Explanations of Neural Activations with MDL-SAEscited by 22
  23. 2022The Rate-Distortion-Perception Tradeoff: The Role of Common Randomnesscited by 21
  24. 2024RDEIC: Accelerating Diffusion-Based Extreme Image Compression with Relay Residual Diffusioncited by 21
  25. 2024The Complexity Dynamics of Grokkingcited by 21
  26. 2025DLF: Extreme Image Compression with Dual-generative Latent Fusioncited by 20
  27. 2022Flexible Neural Image Compression via Code Editingcited by 19
  28. 2024Bidirectional Consistency Modelscited by 18
  29. 2022A Coding Framework and Benchmark towards Low-Bitrate Video Understandingcited by 17
  30. 2024Controlling Rate, Distortion, and Realism: Towards a Single Comprehensive Neural Image Compression Modelcited by 16
  31. 2024PerCo (SD): Open Perceptual Compressioncited by 16
  32. 2024Epsilon-VAE: Denoising as Visual Decodingcited by 16
  33. 2024Once-for-All: Controllable Generative Image Compression with Dynamic Granularity Adaptationcited by 15
  34. 2025FlowDec: A flow-based full-band general audio codec with high perceptual qualitycited by 15
  35. 2021Learned Image Compression for Machine Perceptioncited by 14
  36. 2023M2T: Masking Transformers Twice for Faster Decodingcited by 14
  37. 2026OAT: Ordered Action Tokenizationcited by 14
  38. 2022Optimally Controllable Perceptual Lossy Compressioncited by 13

50 most cited of 207 All 207 (JSON)