Pith. sign in

REVIEW 3 major objections 5 minor 43 references

Efficient Progressive Image Compression with Variance-aware Masking

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that ordering residual-latent elements by predicted variance turns a single trained codec into a competitive progressive stream at roughly half the decoding cost.

desk verdict Solid efficiency gains; the variance mask is never isolated, so the central novelty is unproven. read the letter →

arxiv 2411.10185 v3 pith:LIJAUZSG submitted 2024-11-15 cs.CV

Alberto Presta , Enzo Tartaglione , Attilio Fiandrotti , Marco Grangetto , Pamela Cosman This is my paper · ORCID
classification cs.CV
keywords progressiveimagecompressionfine-grainedscalabilitylearnedvariance-awaremaskingresiduallatentrepresentationrateenhancementmoduleschannel-wiseentropymodelrate-distortionoptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proposes a learned image codec that sends a low-quality base representation first, then refines it by transmitting pieces of the residual difference between the base and a high-quality representation. A masking policy ranks residual elements by the standard deviation the entropy model predicts for them and sends the highest-variance elements first, replacing omitted elements with the predicted mean so the decoder can reconstruct at any intermediate quality. The paper reports that this simple, parameter-free mask yields rate-distortion performance competitive with the leading progressive trit-plane codec while roughly halving GFLOPs, decoding time, and parameter count.

What carries the argument

The central object is the variance-aware mask $m^q_i = M(\sigma^t_i, q)$, a binary tensor computed by thresholding the predicted standard deviation $\sigma^t_i$ at the $(100-q)$-th percentile of the slice's values. The same mask can be computed at both the encoder and the decoder because $\sigma^t_i$ is part of the entropy model, so the encoder skips masked residual elements without signaling their positions. The other load-bearing pieces are the residual latent $r^t_i = y^t_i - \hat{y}^b_i$, which concentrates the information added as quality grows, and the Progressive Channel-wise Entropy Estimation Module (PCEEM), which predicts $\mu^t_i$ and $\sigma^t_i$ slice by slice; the Rate Enhancement Modules refine those predictions at a few checkpoint qualities. Together these mechanisms turn a single trained model into a bitstream that can be truncated at any quality $q$.

What would settle it

A decisive experiment would compare the sigma-based mask against a random mask or a mask ordered by the magnitude of the residual under identical bitrates; if the rate-distortion curves overlap, variance is not what carries the progressive gain.

Watch

Extended reading notes

Core claim

The paper's central claim is that a single learned codec can deliver fine-grained progressive compression if the residual latent $\hat{r}^q_i = Q(r^t_i - \mu^t_i) \otimes m^q_i + \mu^t_i$ is masked according to a percentile threshold on $\sigma^t_i$, the standard deviation predicted by the channel-wise entropy model. Elements whose predicted variance falls below the $(100-q)$-th percentile are not encoded and are replaced by $\mu^t_i$, while all other elements are transmitted; because the mask is a deterministic function of $\sigma^t_i$, the decoder can reconstruct it without side information. The paper shows that with decoder refinement and up to three checkpoint-based Rate Enhancement Modules, this scheme matches the rate-distortion performance of the leading progressive method on Kodak and CLIC and trails only slightly on JPEG-AI, while using about 40 percent of the decoder parameters and roughly 40 percent of the decoding GFLOPs.

Load-bearing premise

The argument assumes that ranking residual elements by the entropy model's predicted standard deviation is a valid ordering of their importance for reconstruction quality, so that sending high-variance elements first is always the best use of bits.

Editorial extensions

If this is right

  • One trained model can serve arbitrary bitrate targets by signaling only $q$, replacing per-rate models or retraining for each quality.
  • The reported BD-Rate on Kodak is $-1.05\%$ and on CLIC $-0.75\%$ relative to the leading progressive codec, with $+0.47\%$ on JPEG-AI.
  • Decoding GFLOPs drop from about 2012 to about 788 on Kodak, and decoder parameters drop from 399M to 90.8M.
  • At least two REM checkpoints are needed to beat the reference method on Kodak; decoder refinement alone yields $+1.24\%$ BD-Rate, and no refinement yields $+5.53\%$.
  • Complexity savings shrink at higher qualities because checkpoint latent representations must be decoded before the target quality can be reached.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The variance-ranking assumption is testable: if a random mask matched the sigma-based mask in rate-distortion, the entire quality-progression argument would weaken, since the paper reports no such ablation.
  • Because the mask is a deterministic percentile function, the same scheme could be applied in variable-rate settings or to a base layer that is itself scalable, a direction the paper names as future work.
  • The checkpoint REMs divide the bitrate range into subranges; a continuous context model that conditions on all previously decoded lower-quality content could replace them and extend the usable bitrange.
  • The residual-mask design resembles residual video coding, so the method may carry over to inter-frame compression, though the paper does not evaluate that.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes a learned progressive image compression method. A base and a top latent are extracted for an image, and a residual latent is formed as their element-wise difference. Fine-grained scalability is obtained by a masking policy that ranks residual elements by the predicted standard deviation from a channel-wise entropy model and encodes only elements above a quality-dependent percentile threshold; omitted elements are replaced by the entropy model mean at the decoder. The paper also introduces slice-wise Rate Enhancement Modules (REMs) that refine entropy parameter estimates using already decoded checkpoint representations. Experiments on Kodak, JPEG-AI, and CLIC report rate-distortion performance competitive with the trit-plane method of Jeon et al., while roughly halving GFLOPs, decoding time, and parameter count. Source code is made available.

Significance. If the variance-based ranking is genuinely useful, the method is a clean and lightweight way to obtain arbitrary truncation points from a single trained model, and the reported computational savings are substantial. The paper's strengths include a simple, parameter-free masking procedure, a reproducible implementation, and a systematic complexity comparison. However, the central novelty -- the sigma-based importance ranking -- is not isolated experimentally: no ablation compares it with random or alternative masks, and the reported RD differences relative to Jeon et al. are so small that without error bars the main claims are not yet demonstrated. The contribution is therefore plausible but insufficiently supported in its current form.

major comments (3)
  1. [Section 3.3, Algorithm 1; Table 3; Fig. 10] The paper's central claim is that sorting residual elements by predicted sigma_t_i gives an importance ordering that makes truncation RD-efficient. This is never tested against any alternative ordering: every configuration in Table 3 and Fig. 9 uses the variance mask, and Fig. 10 is qualitative only. The reported BD-Rate improvements could therefore come entirely from the REMs and decoder refinement rather than from the masking policy. Please add a matched-rate ablation comparing the variance mask with (a) a random mask of the same cardinality, (b) a magnitude-based mask using |r_t_i - mu_t_i|, and (c), if feasible, an oracle ordering based on actual reconstruction error contribution, while keeping REMs, decoder refinement, and rate points fixed. Report BD-Rate/BD-PSNR for each variant.
  2. [Table 1 and Fig. 5] The rate-distortion differences to Jeon et al. are very small (BD-Rate -1.05, +0.47, -0.75 on Kodak, JPEG-AI, and CLIC; BD-PSNR within 0.04 dB). No error bars or multiple seeds are reported, so the statements in Sec. 4.2 that the method 'outperformed' Jeon et al. at low/medium bitrates and is 'competitive' are not statistically supported. Please train and evaluate at least three seeds and report means with standard deviations or confidence intervals for BD-Rate, BD-PSNR, and the RD curves.
  3. [Section 4.1 and Table 3] The REM checkpoint qualities q_bar = {0.5, 7.5, 20} are described as 'empirically selecting' values that correspond to approximately 5%, 40%, and 65% of the Kodak bitstream, and Kodak is also one of the evaluation datasets. This selection procedure can inflate the reported Kodak BD-Rate and makes the comparison less clean. Please select checkpoints on a validation set disjoint from all test datasets, or report the sensitivity of BD-Rate to checkpoint placement on all three datasets, so that the reported gains can be attributed to the method rather than to test-set-informed tuning.
minor comments (5)
  1. [Abstract and Section 1] Tense and grammar should be fixed: 'We also introduced Rate Enhancement Modules' should be 'We also introduce...', and 'the follows' should be 'the following'. In Section 1, the sentence 'allowing to truncate it at' is incomplete and should state the truncation point or quality.
  2. [Figure 3 caption] The word 'Chuck' appears to be a typo for 'Chunk' or 'Slice'.
  3. [Table 2] Units for GPU/CPU decode and encode times are missing (presumably seconds), and the GPU decode time column uses inconsistent precision (1.15 versus 2.018). Please make units and precision uniform.
  4. [Figure 10] The panels labeled 'std at q=10' and 'std at q=50' cannot depend on q, because sigma_t_i is independent of q (Sec. 3.2). Relabel them as the residual standard deviation or describe what is actually shown to avoid implying that the entropy model's sigma changes with quality.
  5. [Section 4.2] The claim that the method outperforms Jeon et al. for low and medium bit rates is not quantified. Please specify the rate interval and report BD-Rate or average PSNR gain over that subrange.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: all central claims are evaluated against external benchmarks, and the variance-ranking heuristic is an untested assumption rather than a derived result.

full rationale

The paper's central derivation chain is empirical rather than definitional. The residual latent is defined as the element-wise difference between top and base latents (Sec. 3.2), and the masking policy ranks residual elements by predicted standard deviation (Sec. 3.3, Alg. 1). This ranking is justified by the heuristic that larger-variance elements are more likely to cause greater reconstruction error; this is an unvalidated assumption, not a circular reduction. No equation equates the predicted output to the fitted input: the rate-distortion results in Fig. 5 and Table 1 are measured against external datasets (Kodak, CLIC, JPEG-AI) and compared with independent methods (Jeon et al., Lee et al., Lu et al., JPEG2000). The REM checkpoints are chosen empirically from the Kodak bitrate distribution, which is a data-selection concern, not circularity, because the reported BD-Rate is still measured on held-out images rather than constructed from the checkpoint choices. The only self-citation is Presta et al. [32] in related work, and it is not load-bearing for any claim in the paper. The core novelties—masking without signaling overhead, REM-based entropy refinement, and complexity reduction—are all evaluated by external measurement and would be falsifiable if the variance ranking were no better than random. Thus, the paper shows no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entity; REMs and the mask are architectural components, not independent evidence-bearing objects. The ledger instead shows tuning choices and assumptions: lambda weights, three checkpoint qualities, and the sigma-is-importance assumption.

free parameters (3)
  • Lagrangian weights lambda_b, lambda_t = 5e-3, 5e-2
    Set in Sec. 4.1; they fix the base and top operating points and shape the whole rate-distortion trade-off.
  • REM checkpoint qualities = 0.5, 7.5, 20
    Selected in Sec. 4.1 from the Kodak bitstream distribution; the reported Kodak BD-Rate depends on this choice.
  • Number of REMs = 3
    Chosen empirically in Sec. 4.1 to balance performance and complexity; Table 3 shows the RD gain emerges only with two or more REMs.
assumptions (4)
  • domain assumption Residual latent elements are independent Gaussians with entropy parameters (mu, sigma) from a hyperprior and channel-wise model.
    Used in Sec. 3.2, Eq. 2 and Eq. 3, and in ANS coding; inherited from Ballé et al. and Minnen and Singh.
  • ad hoc to paper Predicted sigma ranking equals residual importance for reconstruction.
    Stated in Sec. 3.3; no ablation against random or alternative masks is provided.
  • domain assumption Masked residual values can be replaced by mu_t_i without decoder mismatch or meaningful quality loss.
    Eq. 3 requires both sides to agree on the mask and mean, which holds, but the reconstruction quality of skipped elements is assumed.
  • domain assumption Checkpoint REMs trained with the rest of the network frozen improve rate without destabilizing the decoder.
    Sec. 3.5 and the ablation in Fig. 9 and Table 3; depends on training-order choices and empirical checkpoint positions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Progressive Image Compression with Variance-aware Masking." pith.science (2026). https://pith.science/paper/LIJAUZSG

@misc{pith2026241110185,
  author       = {Pith},
  title        = {Pith review of: Efficient Progressive Image Compression with Variance-aware Masking},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LIJAUZSG}},
  note         = {Machine review of arXiv:2411.10185}
}
read the original abstract

Learned progressive image compression is gaining momentum as it allows improved image reconstruction as more bits are decoded at the receiver. We propose a progressive image compression method in which an image is first represented as a pair of base-quality and top-quality latent representations. Next, a residual latent representation is encoded as the element-wise difference between the top and base representations. Our scheme enables progressive image compression with element-wise granularity by introducing a masking system that ranks each element of the residual latent representation from most to least important, dividing it into complementary components, which can be transmitted separately to the decoder in order to obtain different reconstruction quality. The masking system does not add further parameters nor complexity. At the receiver, any elements of the top latent representation excluded from the transmitted components can be independently replaced with the mean predicted by the hyperprior architecture, ensuring reliable reconstructions at any intermediate quality level. We also introduced Rate Enhancement Modules (REMs), which refine the estimation of entropy parameters using already decoded components. We obtain results competitive with state-of-the-art competitors, while significantly reducing computational complexity, decoding time, and number of parameters.

Figures

Figures reproduced from arXiv: 2411.10185 by the authors.

Figure 1
Figure 1. Compression results for three different qualities, which [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our proposed architecture. Green and red [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Progressive channel wise entropy estimation model ( [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Blueprint of REM for a fixed checkpoint quality [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Rate-distortion performance of our method compared with progressive image compression algorithms: Jeon [ [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Kodim14 reconstruction from Kodak dataset by different codecs: Proposed, Jeon [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: RD curve for some fixed-rate models on Kodak: [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: GFLOPs and decoding time complexity on NVIDIA [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Results on Kodak considering different configurations. [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 10
Figure 10. Figure 10: Original image (a), base latent channel and estimate [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 37 canonical work pages

  1. [1]

    Jpeg-ai test images. 2023. 6

    work page 2023

  2. [2]

    Variational image compression with a scale hyperprior

    Johannes Ball ´e, , et al. Variational image compression with a scale hyperprior. arXiv preprint arXiv:1802.01436, 2018. 2, 3, 7

    arXiv 2018

  3. [3]

    End-to-end optimized image compression

    Johannes Ball ´e, Valero Laparra, and Eero P Simoncelli. End-to-end optimized image compression. arXiv preprint arXiv:1611.01704, 2016. 2

    arXiv 2016

  4. [4]

    Compressai: a pytorch library and eval- uation platform for end-to-end compression research

    Jean B ´egaint et al. Compressai: a pytorch library and eval- uation platform for end-to-end compression research. arXiv preprint arXiv:2011.03029, 2020. 6

    arXiv 2011

  5. [5]

    Overview of the versatile video coding (vvc) standard and its applications

    Benjamin Bross et al. Overview of the versatile video coding (vvc) standard and its applications. IEEE Transactions on Circuits and Systems for Video Technology, 2021. 1, 3

    work page 2021

  6. [6]

    A novel deep progressive image compres- sion framework

    Chunlei Cai et al. A novel deep progressive image compres- sion framework. In Picture Coding Symposium , 2019. 2, 3

    work page 2019

  7. [7]

    Learned image compression with dis- cretized gaussian mixture likelihoods and attention modules

    Zhengxue Cheng et al. Learned image compression with dis- cretized gaussian mixture likelihoods and attention modules. In CVPR, 2020. 2, 7

    work page 2020

  8. [8]

    Asymmetric gained deep image compression with continuous rate adaptation

    Ze Cui et al. Asymmetric gained deep image compression with continuous rate adaptation. In CVPR, 2021. 1

    work page 2021

Show all 43 references
  1. [9]

    Drasic: Distributed recurrent autoencoder for scalable image compression

    Enmao Diao et al. Drasic: Distributed recurrent autoencoder for scalable image compression. In DCC. IEEE, 2020. 1, 2

    2020

  2. [10]

    Asymmetric numeral systems: entropy coding combining speed of huffman coding with compression rate of arithmetic coding

    Jarek Duda. Asymmetric numeral systems: entropy coding combining speed of huffman coding with compression rate of arithmetic coding. arXiv preprint arXiv:1311.2540, 2013. 4

    2013 arXiv

  3. [11]

    Kodak lossless true color image suite

    Rich Franzen. Kodak lossless true color image suite. source: http://r0k. us/graphics/kodak, 4(2), 1999. 6

    1999

  4. [12]

    Towards conceptual com- pression

    Karol Gregor, Frederic Besse, Danilo Jimenez Rezende, Ivo Danihelka, and Daan Wierstra. Towards conceptual com- pression. NeurIPS, 29, 2016. 2

    2016

  5. [13]

    Checkerboard context model for efficient learned image compression

    Dailan He et al. Checkerboard context model for efficient learned image compression. In CVPR, June 2021. 2

    2021

  6. [14]

    Long short-term memory

    Sepp Hochreiter and J ¨urgen Schmidhuber. Long short-term memory. Neural computation, 9(8), 1997. 2

    1997

  7. [15]

    Context-based trit-plane coding for progressive im- age compression

    Seungmin Jeon, Kwang Pyo Choi, Youngo Park, and Chang- Su Kim. Context-based trit-plane coding for progressive im- age compression. In CVPR, 2023. 2, 3, 5, 6, 8

    2023

  8. [16]

    Neural image compression using masked sparse visual representation

    Wei Jiang et al. Neural image compression using masked sparse visual representation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2024. 2

    2024

  9. [17]

    Improved lossy image compression with priming and spatially adaptive bit rates for recurrent networks

    Nick Johnston et al. Improved lossy image compression with priming and spatially adaptive bit rates for recurrent networks. In CVPR, 2018. 2

    2018

  10. [18]

    Variable-rate learned image com- pression with multi-objective optimization and quantization- reconstruction offsets

    Fatih Kamisli et al. Variable-rate learned image com- pression with multi-objective optimization and quantization- reconstruction offsets. In DCC. IEEE, 2024. 1

    2024

  11. [19]

    Auto-encoding varia- tional bayes

    Diederik P Kingma and Max Welling. Auto-encoding varia- tional bayes. arXiv preprint arXiv:1312.6114, 2013. 2

    2013 arXiv

  12. [20]

    The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale

    Alina Kuznetsova et al. The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale. IJCV, 128(7), 2020. 6

    2020

  13. [21]

    Selective compression learning of latent representations for variable- rate image compression

    Jooyoung Lee, Seyoon Jeong, and Munchurl Kim. Selective compression learning of latent representations for variable- rate image compression. In NeurIPS, 2022. 1

    2022

  14. [22]

    Dpict: Deep progressive image compression using trit-planes

    Jae-Han Lee, Seungmin Jeon, Kwang Pyo Choi, Youngo Park, and Chang-Su Kim. Dpict: Deep progressive image compression using trit-planes. In CVPR, 2022. 1, 2, 3, 6, 8

    2022

  15. [23]

    Overview of fine granularity scalability in mpeg- 4 video standard

    Weiping Li. Overview of fine granularity scalability in mpeg- 4 video standard. IEEE Transactions on circuits and systems for video technology, 11(3), 2001. 1

    2001

  16. [24]

    Deep learning-based video coding: A review and a case study

    Dong Liu et al. Deep learning-based video coding: A review and a case study. ACM Computing Surveys, 53(1), 2020. 4

    2020

  17. [25]

    Non-local attention optimized deep image compression

    Haojie Liu et al. Non-local attention optimized deep image compression. arXiv preprint arXiv:1904.09757, 2019. 2

    1904 arXiv

  18. [26]

    Learned image compression with mixed transformer-CNN architectures

    Jinming Liu et al. Learned image compression with mixed transformer-CNN architectures. In CVPR, 2023. 1, 2, 3, 6, 7

    2023

  19. [27]

    Progressive neural image compression with nested quantization and latent ordering

    Yadong Lu, Yinhao Zhu, Yang Yang, Amir Said, and Taco S Cohen. Progressive neural image compression with nested quantization and latent ordering. In ICIP, 2021. 1, 3, 6

    2021

  20. [28]

    Recurrent neural net- works

    Larry R Medsker, Lakhmi Jain, et al. Recurrent neural net- works. Design and Applications, 5(64-67):2, 2001. 2

    2001

  21. [29]

    Joint autoregressive and hierarchical priors for learned im- age compression

    David Minnen, Johannes Ball ´e, and George D Toderici. Joint autoregressive and hierarchical priors for learned im- age compression. NeurIPS, 31, 2018. 2, 7

    2018

  22. [30]

    Channel-wise autoregres- sive entropy models for learned image compression

    David Minnen and Saurabh Singh. Channel-wise autoregres- sive entropy models for learned image compression. InICIP,

  23. [31]

    Advances in scalable video coding

    J-R Ohm. Advances in scalable video coding. Proceedings of the IEEE, 93(1), 2005. 1

    2005

  24. [32]

    A differentiable entropy model for learned image compression

    Alberto Presta et al. A differentiable entropy model for learned image compression. In ICIAP. Springer, 2023. 2

    2023

  25. [33]

    The jpeg 2000 still image compression standard

    Athanassios Skodras, Charilaos Christopoulos, and Touradj Ebrahimi. The jpeg 2000 still image compression standard. IEEE Signal processing magazine, 18(5), 2001. 1, 6

    2000

  26. [34]

    Gabic: Graph-based attention block for image compression

    Gabriele Spadaro et al. Gabic: Graph-based attention block for image compression. In ICIP, 2024. 2

    2024

  27. [35]

    Lossy image compression with compressive autoen- coders

    Lucas Theis, Wenzhe Shi, Andrew Cunningham, and Ferenc Husz´ar. Lossy image compression with compressive autoen- coders. In ICLR, 2017. 2

    2017

  28. [36]

    Variable rate image compression with recurrent neural networks

    George Toderici et al. Variable rate image compression with recurrent neural networks. ICLR, 2015. 2

    2015

  29. [37]

    Full resolution image compression with recurrent neural networks

    George Toderici et al. Full resolution image compression with recurrent neural networks. In CVPR, 2017. 2

    2017

  30. [38]

    Workshop and challenge on learned image compression (clic2020)

    George Toderici et al. Workshop and challenge on learned image compression (clic2020). In CVPR, 2020. 6

    2020

  31. [39]

    The jpeg still picture compression stan- dard

    Gregory K Wallace. The jpeg still picture compression stan- dard. Communications of the ACM, 34(4), 1991. 1

    1991

  32. [40]

    Evc: To- wards real-time neural image compression with mask decay

    Guo-Hua Wang, Jiahao Li, Bin Li, and Yan Lu. Evc: To- wards real-time neural image compression with mask decay. In ICLR, 2023. 3

    2023

  33. [41]

    Enhanced invertible encoding for learned image compression

    Yueqi Xie et al. Enhanced invertible encoding for learned image compression. In ACM MM, 2021. 2

    2021

  34. [42]

    End-to-end optimized image compression with attention mechanism

    Lei Zhou et al. End-to-end optimized image compression with attention mechanism. In CVPRW, 2019. 2

    2019

  35. [43]

    The devil is in the details: Window-based attention for image compression

    Renjie Zou, Chunfeng Song, and Zhaoxiang Zhang. The devil is in the details: Window-based attention for image compression. In CVPR, 2022. 1, 2, 3, 6, 7

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.