Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

Learned Image Compression with Hierarchical Progressive Context Modeling

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that coding latents coarse-to-fine with context carried across scales gives learned image compression state-of-the-art rate-distortion performance at reduced complexity.

desk verdict Solid architecture paper with a credible RD gain and clean ablations, but the Base model's SOTA claim is a single-seed, sub-0.3pp margin over MLIC++ with no variance reported. read the letter →

arxiv 2507.19125 v1 pith:EVMD45UM submitted 2025-07-25 eess.IV cs.CVcs.MM

Yuqi Li , Haotian Zhang , Li Li , Dong Liu This is my paper · ORCID
classification eess.IVcs.CVcs.MM
keywords learnedimagecompressionentropycodingcontextmodelinghierarchicalscheduleprogressivefusionautoregressivemodelrate-distortionoptimizationlong-rangedependency
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

The paper tries to establish that the expensive part of learned image compression, modeling the probability of every latent symbol so entropy coding uses as few bits as possible, can be made cheaper and more accurate if the latents are coded in a coarse-to-fine order. Instead of paying for global attention over the full latent map, the proposed HPCM codes a 4x-downsampled sub-latent first, then progressively fills in finer scales. A cross-attention fusion step carries context accumulated from earlier coding steps into the next step. If the claim holds, learned codecs can get the gains of long-range context modeling without the quadratic complexity that usually comes with it.

What carries the argument

Three-scale coarse-to-fine latent coding with progressive context fusion. The quantized latents $\hat{\mathbf{y}}$ are split into $\hat{\mathbf{y}}^{S_1}$ (a 4x downsampled version), $\hat{\mathbf{y}}^{S_2}$, and $\hat{\mathbf{y}}^{S_3}$, coded with 2, 3, and 6 steps respectively. After each scale is coded, it is filled back into the next scale, so the context network sees long-range structure before fine detail. The progressive context fusion module maintains a context state $C_i$ per coding step, updates it by cross-attention with the entropy-parameter state $\psi_i$, and propagates it across scales; this carries diverse context from past coding steps into the current estimate of the generalized Gaussian entropy parameters.

What would settle it

Train a variant that replaces the fixed hierarchical schedule with a content-adaptive ordering, or with the same progressive fusion but a single-scale context model, keeping everything else identical; if BD-Rate stays within about 0.5% of HPCM-Base, the hierarchical schedule is not the source of the gain. Alternatively, evaluate the released model on 4K or screen-content images, since the schedule was tuned and validated only at 512x768 and 1200x1200; a clear complexity or rate-distortion regression there would show the schedule does not transfer.

Watch

Extended reading notes

Core claim

The central claim is that a hierarchical progressive context model (HPCM) gives state-of-the-art rate-distortion performance while using substantially less computation than recent transformer-based context models. The method partitions the quantized latent representation into three scales and codes them sequentially from the smallest to the largest, so long-range dependencies are captured cheaply at the coarse scale and short-range details are refined at the fine scale. Across eight channel groups the partition pattern varies, mixing spatial and channel context. A progressive context fusion module accumulates the entropy-model state from previous coding steps through cross-attention, then propagates it across scales by filling coarse context into the finer scale and merging it with the hyperprior. On Kodak, HPCM-Large reports 19.19% BD-Rate savings over VTM-22.0 and HPCM-Base reports 15.31%, outperforming MLIC++ and FLIC; the paper attributes the gain specifically to the hierarchical schedule and the progressive fusion.

Load-bearing premise

The central assumption is that a hand-designed schedule of three scales, eight channel-group partition patterns, and fixed step counts (2, 3, 6) is the right way to spend compute; if a simpler or content-adaptive schedule yielded the same rate-distortion gains, the contribution would reduce to the fusion module.

Editorial extensions

If this is right

  • Long-range context modeling in learned codecs no longer needs full-resolution global attention: the same dependency coverage comes from first coding a coarse sub-latent, with complexity around 918 to 1261 kMACs/pixel.
  • A codec that codes coarse structure first will allocate compute where it pays off: the ablation reports that raising the fine-scale steps from (2,3,6) to (2,3,12) buys about 2.55% more BD-Rate savings at much higher cost, while adding coarse-scale steps buys almost nothing.
  • Sharing context-model weights across coding steps keeps the parameter count near 68.5M, while non-shared weights add about 121M parameters for essentially identical performance.
  • The entropy-model state itself is a useful context carrier: removing fusion entirely costs 4.71% in BD-Rate, while reusing only the previous state recovers most of the loss, showing the cross-attention fusion closes the remaining gap.
  • The released model makes the hierarchical progressive context design directly reproducible and testable on other datasets and resolutions.

Reading between the lines

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

  • If the coding schedule were content-adaptive, choosing the number of scales and per-scale step counts from the image statistics, the same fusion mechanism might push the gains further, since the fixed (2,3,6) allocation is only one operating point and the ablations show diminishing returns at the coarse scale.
  • The coarse-to-fine order already produces an embedded bitstream, so HPCM could be extended to progressive or quality-scalable transmission, where a coarse reconstruction is decoded first and the finer scales refine it.
  • The cross-attention fusion's attention maps concentrate on high-bitrate texture regions, suggesting a testable loop where the fusion module predicts where future coding steps should be spent.
  • The hierarchical schedule resembles scalable wavelet coding, so coupling HPCM with a wavelet-like transform might let the codec allocate context modeling effort across frequency bands as well as scales.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes HPCM, a learned image compression method built around a hierarchical progressive context model. The latent tensor is partitioned into three scales and coded sequentially from the smallest scale to the largest, so that long-range dependencies are modeled efficiently at low resolution before short-range detail is added. A progressive context fusion module uses cross-attention to accumulate context across coding steps and scales, and structural improvements plus a shared-parameter design reduce complexity. On Kodak, CLIC Pro Valid, and Tecnick, the authors report BD-Rate savings over VTM-22.0 of 15.31%, 14.23%, and 18.16% for HPCM-Base and 19.19%, 18.37%, and 22.20% for HPCM-Large, claiming state-of-the-art rate-distortion performance and a better complexity trade-off than existing learned codecs. The code is released, and the paper includes extensive ablations of the coding schedule, fusion mechanism, step allocation, number of stages, and parameter sharing.

Significance. If the reported results hold, HPCM is a meaningful contribution to learned image compression: it reduces kMACs/pixel by roughly 28% relative to MLIC++ while matching or slightly improving BD-Rate in the Base configuration, and the Large configuration improves BD-Rate by ~4 percentage points at comparable complexity. The paper is strong on experimental discipline: entropy-model ablations are run under a fixed transform, the coding-step allocation is tested in four configurations, shared-parameter effects are quantified, and the code is public. The main fragility is that the Base model's state-of-the-art claim rests on BD-Rate margins of 0.16–0.26 percentage points over MLIC++, which is below typical seed-to-seed variation in learned codecs; without uncertainty quantification, that specific claim is not fully supported. The Large model's margin is much larger and supports the core contribution.

major comments (3)
  1. [Section 4.2, Table 1] The state-of-the-art claim for HPCM-Base is not robust to training stochasticity. The BD-Rate advantage over MLIC++ is 0.16 percentage points on Kodak, 0.18 on CLIC Pro Valid, and 0.26 on Tecnick, yet the paper reports no standard deviations, confidence intervals, or multiple-seed results. In learned image compression, single-seed differences smaller than about one BD-Rate point are often within run-to-run variation. The abstract attributes state-of-the-art performance to 'our method' generally, and the Base model's complexity-balance contribution depends on this small gap. The authors should either provide uncertainty quantification (e.g., results from three training seeds with mean and standard deviation) or restrict the state-of-the-art claim to the Large model, which has a substantially larger margin.
  2. [Section 3.2, Tables 2 and D] The hierarchical coding schedule is entirely hand-designed: three scales, eight channel-group-specific partition patterns, and the fixed step allocation (2,3,6). The ablations in Table 2 show that varying the allocation changes BD-Rate by -2.55% to +2.39%, so this choice is load-bearing for the claimed complexity-performance trade-off. However, all evaluations are conducted on 512x768 and 1200x1200 images, and there is no evidence about how the schedule transfers to higher resolutions (e.g., 4K) or to different content types. The paper should either test the schedule at higher resolutions or, if that is infeasible, discuss the expected sensitivity of the design choices to input resolution.
  3. [Section 4.2, Related Work] The comparison set used to support the state-of-the-art claim is a selection of methods with public code, but several recent methods cited in the related work are not evaluated, including GroupedMixer [23] and the method of Kim et al. [20]. If any of these methods achieves better BD-Rate or a better complexity trade-off, the abstract's claim may not hold. The authors should either include these methods in Table 1 (where code is available) or provide a clear justification for their exclusion.
minor comments (6)
  1. [Tables 2 and 3] The headers of Tables 2 and 3 contain a typo: 'Abaltion' should be 'Ablation'.
  2. [Figure 9] In the caption and image labels of Figure 9, 'HCPM-Large' and 'HCPM-Base' should be 'HPCM-Large' and 'HPCM-Base'.
  3. [Section 4.1] The training dataset is referred to as 'Flicker2W'; this should likely be 'Flickr2W' (the standard Flickr 2W dataset).
  4. [Section 3.2] The eight channel-group-specific partition patterns are only fully described in the supplementary figure; a brief textual summary in the main text would help readers who do not consult the supplement.
  5. [Table 1] The MambaVC row has missing entries for CLIC Pro Valid and Tecnick; this is acceptable, but a footnote explaining that the results are unavailable (e.g., code or model not released for those datasets) would improve clarity.
  6. [Supplementary Figure F] The panels in Figure F are dense; enlarging the panels or annotating the step numbers more visibly would make the alternative coding schedules easier to follow.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: benchmark claims rest on external anchors; self-citations are design references, not load-bearing inputs.

full rationale

The paper's central claims are empirical: HPCM-Base/Large achieve the reported BD-Rate gains over VTM-22.0 and open-source learned codecs on Kodak, CLIC Pro Valid, and Tecnick (Table 1, Figs. 5-7). The hierarchical coding schedule and progressive context fusion are architectural proposals evaluated by ablations (Tables 2 and 3, supplement Tables D and E) against alternative schedules and fusion variants; no equation defines the reported performance as a consequence of a fitted parameter or of the proposed components by construction. The few self-citations ([28] for the octree coding order and [49] for the generalized Gaussian entropy model) supply design choices or modeling assumptions, but the SOTA claim stands or falls on external benchmarks and ablations regardless of those citations, so they are not load-bearing. The paper also notes in Sec. F that coding time is implementation-dependent and therefore emphasizes kMACs/pixel; that is an honest scope limitation, not a circularity. Concerns such as the absence of variance estimates for the small BD-Rate margins versus MLIC++ or the transferability of the hand-tuned (2,3,6) schedule are correctness/robustness risks, not circularity.

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

The central claim depends on a set of hand-tuned design choices (coding step allocation, stage count, partition patterns) and on standard domain assumptions of learned compression. No invented entities are introduced. The effective free parameters are the architecture hyperparameters and coding order, which are selected via ablations rather than derived.

free parameters (3)
  • Coding step allocation (S1, S2, S3) = (2, 3, 6)
    Chosen by hand based on ablations; affects BD-Rate vs kMACs trade-off. Ablation (2,3,12) gives 2.55% better BD-Rate at 1427.91 kMACs/pixel, so the default is a deliberate trade-off.
  • Number of hierarchical stages = 3
    Selected via ablation (Table D); 3-stage gives lower kMACs and better BD-Rate than 1, 2, or 4 stages.
  • Channel-group partition patterns = 8 patterns
    Hand-designed to enable spatial-channel interaction; no automated search.
assumptions (4)
  • domain assumption The generalized Gaussian model with shape beta fixed at 1.5 adequately models latent distributions.
    Adopted from prior work [49]; if the model is misspecified, entropy estimates degrade.
  • domain assumption The learned compression framework (analysis/synthesis transforms, hyperprior) is a valid basis for rate-distortion optimization.
    Standard in the field; not re-derived in this paper.
  • domain assumption VTM-22.0 is a fair and stable anchor for BD-Rate calculation.
    Use of a traditional codec anchor is conventional, but BD-Rate values depend on the anchor version.
  • domain assumption Flickr2W training data is representative and sufficient for learning the entropy model.
    Standard training set in prior work [34].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learned Image Compression with Hierarchical Progressive Context Modeling." pith.science (2026). https://pith.science/paper/EVMD45UM

@misc{pith2026250719125,
  author       = {Pith},
  title        = {Pith review of: Learned Image Compression with Hierarchical Progressive Context Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EVMD45UM}},
  note         = {Machine review of arXiv:2507.19125}
}
read the original abstract

Context modeling is essential in learned image compression for accurately estimating the distribution of latents. While recent advanced methods have expanded context modeling capacity, they still struggle to efficiently exploit long-range dependency and diverse context information across different coding steps. In this paper, we introduce a novel Hierarchical Progressive Context Model (HPCM) for more efficient context information acquisition. Specifically, HPCM employs a hierarchical coding schedule to sequentially model the contextual dependencies among latents at multiple scales, which enables more efficient long-range context modeling. Furthermore, we propose a progressive context fusion mechanism that incorporates contextual information from previous coding steps into the current step, effectively exploiting diverse contextual information. Experimental results demonstrate that our method achieves state-of-the-art rate-distortion performance and strikes a better balance between compression performance and computational complexity. The code is available at https://github.com/lyq133/LIC-HPCM.

Figures

Figures reproduced from arXiv: 2507.19125 by the authors.

Figure 1
Figure 1. Comparison between BD-rate over VTM-22.0 and end [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Hierarchical coding schedule of HPCM. yˆ is partitioned into three sub-latents at different scales: yˆ S1 , yˆ S2 , and yˆ S3 . These partition strategies vary across different channel groups. We show the partition process of the first and eighth channel groups in this figure. Sequential conditional entropy coding and progressive context modeling are then applied at each coding step. We begin the coding process on y… view at source ↗
Figure 3
Figure 3. Diagram of our proposed Progressive Context Fusion (PCF) method. We take the coding process of [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Illustration of Progressive Context Fusion (PCF) mod [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Rate-distortion curves on Kodak dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Rate-distortion curves on CLIC Pro Valid dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 8
Figure 8. Figure 8: Illustration of methods to obtain the entropy parameter [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: Visualization of reconstructed images of Kodim24 in the Kodak dataset with various learned image compression methods. [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 11
Figure 11. Figure 11: Visualization of (a) Original image Kodim01 from the [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Next-Frame Decoding for Ultra-Low-Bitrate Image Compression with Video Diffusion Priors

    cs.CV 2026-03 conditional novelty 6.5 of 10

    Ultra-low-bitrate image decoding is cast as one-step next-frame prediction from a compact anchor using adapted video diffusion priors, yielding large perceptual bitrate savings versus DiffC.

  2. Cool-chic 5.0: Faster Encoding and Inter-Feature Entropy Modeling for Overfitted Image Compression

    eess.IV 2026-05 unverdicted novelty 6.0 of 10

    Cool-chic 5.0 delivers 11% lower rate than H.266/VVC and matches modern autoencoders like MLIC++ with 250 times lower decoding complexity through an updated decoder architecture and faster optimization for overfitted codecs.

Reference graph

Works this paper leans on

53 extracted references · 45 canonical work pages · cited by 2 Pith papers

  1. [23]

    GroupedMixer: An entropy model with group-wise token-mixers for learned image compres- sion

    Daxin Li, Yuanchao Bai, Kai Wang, Junjun Jiang, Xian- ming Liu, and Wen Gao. GroupedMixer: An entropy model with group-wise token-mixers for learned image compres- sion. IEEE Transactions on Circuits and Systems for Video Technology, 34(10):9606–9619, 2024. 4

    work page 2024

  2. [20]

    Diversify, contextualize, and adapt: Efficient en- tropy modeling for neural image codec

    Jun-Hyuk Kim, Seungeon Kim, Won-Hee Lee, and Dok- wan Oh. Diversify, contextualize, and adapt: Efficient en- tropy modeling for neural image codec. pages 45956–45974,

    work page

  3. [1]

    TESTIMAGES: A large-scale archive for testing visual devices and basic image processing algorithms

    Nicola Asuni and Andrea Giachetti. TESTIMAGES: A large-scale archive for testing visual devices and basic image processing algorithms. In Smart Tools and Apps for Graph- ics - Eurographics Italian Chapter Conference, pages 63–70,

    work page

  4. [2]

    Simoncelli

    Johannes Ball ´e, Valero Laparra, and Eero P. Simoncelli. End-to-end optimized image compression. In International Conference on Learning Representations (ICLR), 2017. 2

    work page 2017

  5. [3]

    Variational image compres- sion with a scale hyperprior

    Johannes Ball ´e, David Minnen, Saurabh Singh, Sung Jin Hwang, and Nick Johnston. Variational image compres- sion with a scale hyperprior. In International Conference on Learning Representations (ICLR), 2018. 1, 2

    work page 2018

  6. [4]

    CompressAI: A PyTorch library and evalua- tion platform for end-to-end compression research

    Jean B ´egaint, Fabien Racap ´e, Simon Feltman, and Akshay Pushparaja. CompressAI: A PyTorch library and evalua- tion platform for end-to-end compression research. arXiv preprint arXiv:2011.03029, 2020. 13

    arXiv 2011

  7. [5]

    F. Bellard. BPG Image Format. http://bellard.org/ bpg/, 2015. 1

    work page 2015

  8. [6]

    Calculation of average psnr differences between rd-curves

    Gisle Bjontegaard. Calculation of average psnr differences between rd-curves. ITU SG16 Doc. VCEG-M33, 2001. 5

    work page 2001

Show all 53 references
  1. [7]

    Sullivan, and Jens-Rainer Ohm

    Benjamin Bross, Ye-Kui Wang, Yan Ye, Shan Liu, Jianle Chen, Gary J. Sullivan, and Jens-Rainer Ohm. Overview of the versatile video coding (VVC) standard and its applica- tions. IEEE Transactions on Circuits and Systems for Video Technology, 31(10):3736–3764, 2021. 1

    2021

  2. [8]

    Gary Chan

    Jierun Chen, Shiu-hong Kao, Hao He, Weipeng Zhuo, Song Wen, Chul-Ho Lee, and S.-H. Gary Chan. Run, don’t walk: Chasing higher FLOPS for faster neural networks. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 12021–12031, 2023. 5, 11

    2023

  3. [9]

    End-to-end learnt image compression via non-local attention optimization and improved context mod- eling

    Tong Chen, Haojie Liu, Zhan Ma, Qiu Shen, Xun Cao, and Yao Wang. End-to-end learnt image compression via non-local attention optimization and improved context mod- eling. IEEE Transactions on Image Processing , 30:3179– 3191, 2021. 2

    2021

  4. [10]

    Learned image compression with discretized Gaussian mixture likelihoods and attention modules

    Zhengxue Cheng, Heming Sun, Masaru Takeuchi, and Jiro Katto. Learned image compression with discretized Gaussian mixture likelihoods and attention modules. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 7936–7945, 2020. 2

    2020

  5. [11]

    Xception: Deep learning with depthwise separable convolutions

    Franc ¸ois Chollet. Xception: Deep learning with depthwise separable convolutions. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pages 1800–1807,

  6. [12]

    Weconvene: Learned im- age compression with wavelet-domain convolution and en- tropy model

    Haisheng Fu, Jie Liang, Zhenman Fang, Jingning Han, Feng Liang, and Guohe Zhang. Weconvene: Learned im- age compression with wavelet-domain convolution and en- tropy model. In European Conference on Computer Vision (ECCV), pages 37–53, 2024. 6, 7, 12, 13

    2024

  7. [13]

    V .K. Goyal. Theoretical foundations of transform coding. IEEE Signal Processing Magazine, 18(5):9–21, 2001. 1

    2001

  8. [14]

    Soft then hard: Rethinking the quantization in neural image compression

    Zongyu Guo, Zhizheng Zhang, Runsen Feng, and Zhibo Chen. Soft then hard: Rethinking the quantization in neural image compression. In International Conference on Machine Learning (ICML), pages 3920–3929, 2021. 2

    2021

  9. [15]

    Checkerboard context model for efficient learned image compression

    Dailan He, Yaoyan Zheng, Baocheng Sun, Yan Wang, and Hongwei Qin. Checkerboard context model for efficient learned image compression. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 14766–14775, 2021. 1, 2

    2021

  10. [16]

    ELIC: Efficient learned image compres- sion with unevenly grouped space-channel contextual adap- tive coding

    Dailan He, Ziming Yang, Weikun Peng, Rui Ma, Hongwei Qin, and Yan Wang. ELIC: Efficient learned image compres- sion with unevenly grouped space-channel contextual adap- tive coding. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 5708–5717, 202...

    2022

  11. [17]

    Fvc: A new frame- work towards deep video compression in feature space

    Zhihao Hu, Guo Lu, and Dong Xu. Fvc: A new frame- work towards deep video compression in feature space. In 2021 IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 1502–1511, 2021. 2

    2021

  12. [18]

    Towards practical real-time neural video compression

    Zhaoyang Jia, Bin Li, Jiahao Li, Wenxuan Xie, Linfeng Qi, Houqiang Li, and Yan Lu. Towards practical real-time neural video compression. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2025, Nashville, TN, USA, June 11-25, 2024, 2025. 2

    2025

  13. [19]

    MLIC++: Linear complex- ity multi-reference entropy modeling for learned image com- pression

    Wei Jiang and Ronggang Wang. MLIC++: Linear complex- ity multi-reference entropy modeling for learned image com- pression. In ICML Workshop, 2023. 1, 2, 6, 7, 12, 13

    2023

  14. [21]

    Kingma and Jimmy Ba

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR), 2015. 5

    2015

  15. [22]

    Kodak lossless true color image suite

    Eastman Kodak. Kodak lossless true color image suite. http://r0k.us/graphics/kodak/. 5

  16. [24]

    Frequency-aware transformer for learned image compression

    Han Li, Shaohui Li, Wenrui Dai, Chenglin Li, Junni Zou, and Hongkai Xiong. Frequency-aware transformer for learned image compression. In International Conference on Learn- ing Representations (ICLR), 2024. 1, 2, 6, 7, 12, 13

    2024

  17. [25]

    Deep contextual video com- pression

    Jiahao Li, Bin Li, and Yan Lu. Deep contextual video com- pression. Advances in Neural Information Processing Sys- tems, 34, 2021. 2

    2021

  18. [26]

    Neural video compression with diverse contexts

    Jiahao Li, Bin Li, and Yan Lu. Neural video compression with diverse contexts. In IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 22616– 22626, 2023. 1, 2, 4, 5, 6

    2023

  19. [27]

    Neural video compression with feature modulation

    Jiahao Li, Bin Li, and Yan Lu. Neural video compression with feature modulation. In IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, CVPR 2024, Seattle, WA, USA, June 17-21, 2024, 2024. 2

    2024

  20. [28]

    Flexible coding or- der for learned image compression

    Yuqi Li, Haotian Zhang, and Dong Liu. Flexible coding or- der for learned image compression. In IEEE International Conference on Visual Communications and Image Process- ing (VCIP), pages 1–5, 2023. 4, 11

    2023

  21. [29]

    Deviation control for learned 9 image compression

    Yuqi Li, Haotian Zhang, Xiaomin Song, Zheng Liu, Huiming Zheng, Li Li, and Dong Liu. Deviation control for learned 9 image compression. In 2024 IEEE International Conference on Visual Communications and Image Processing (VCIP) , pages 1–5, 2024. 2

    2024

  22. [30]

    In-loop filtering via trained look-up tables

    Zhuoyuan Li, Jiacheng Li, Yao Li, Li Li, Dong Liu, and Feng Wu. In-loop filtering via trained look-up tables. In 2024 IEEE International Conference on Visual Communications and Image Processing (VCIP), pages 1–5. IEEE, 2024. 2

    2024

  23. [31]

    Uniformly accelerated motion model for inter pre- diction

    Zhuoyuan Li, Yao Li, Chuanbo Tang, Li Li, Dong Liu, and Feng Wu. Uniformly accelerated motion model for inter pre- diction. In 2024 IEEE International Conference on Visual Communications and Image Processing (VCIP) , pages 1–5. IEEE, 2024

    2024

  24. [32]

    Object segmentation-assisted inter prediction for versatile video coding

    Zhuoyuan Li, Zikun Yuan, Li Li, Dong Liu, Xiaohu Tang, and Feng Wu. Object segmentation-assisted inter prediction for versatile video coding. IEEE Transactions on Broadcast- ing, 2024

    2024

  25. [33]

    USTC-TD: A test dataset and bench- mark for image and video coding in 2020s

    Zhuoyuan Li, Junqi Liao, Chuanbo Tang, Haotian Zhang, Yuqi Li, Yifan Bian, Xihua Sheng, Xinmin Feng, Yao Li, Changsheng Gao, et al. USTC-TD: A test dataset and bench- mark for image and video coding in 2020s. IEEE Transac- tions on Multimedia, 2025. 2

    2025

  26. [34]

    A unified end-to-end framework for efficient deep image compression

    Jiaheng Liu, Guo Lu, Zhihao Hu, and Dong Xu. A unified end-to-end framework for efficient deep image compression. arXiv preprint arXiv:2002.03370, 2020. 5

    2002 arXiv

  27. [35]

    Learned image compression with mixed Transformer-CNN architectures

    Jinming Liu, Heming Sun, and Jiro Katto. Learned image compression with mixed Transformer-CNN architectures. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14388–14397, 2023. 1, 2, 6, 7, 12, 13

    2023

  28. [36]

    Understanding the effective receptive field in deep convolu- tional neural networks

    Wenjie Luo, Yujia Li, Raquel Urtasun, and Richard Zemel. Understanding the effective receptive field in deep convolu- tional neural networks. 2016. 8

    2016

  29. [37]

    End-to-end optimized versatile image compression with wavelet-like transform

    Haichuan Ma, Dong Liu, Ning Yan, Houqiang Li, and Feng Wu. End-to-end optimized versatile image compression with wavelet-like transform. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, 44(3):1247–1263, 2022. 2

    2022

  30. [38]

    M2T: Masking Transformers twice for faster decoding

    Fabian Mentzer, Eirikur Agustson, and Michael Tschannen. M2T: Masking Transformers twice for faster decoding. In IEEE/CVF International Conference on Computer Vision (ICCV), pages 5317–5326, 2023. 4, 11

    2023

  31. [39]

    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. InIEEE International Conference on Image Processing (ICIP), pages 3339–3343, 2020. 1, 2, 6

    2020

  32. [40]

    Joint autoregressive and hierarchical priors for learned image compression

    David Minnen, Johannes Ball ´e, and George Toderici. Joint autoregressive and hierarchical priors for learned image compression. In Advances in Neural Information Process- ing Systems, pages 10794–10803, 2018. 1, 2

    2018

  33. [41]

    Entroformer: A Transformer-based entropy model for learned image compression

    Yichen Qian, Ming Lin, Xiuyu Sun, Zhiyu Tan, and Rong Jin. Entroformer: A Transformer-based entropy model for learned image compression. In International Conference on Learning Representations (ICLR), 2022. 1, 2, 4

    2022

  34. [42]

    Mambavc: Learned visual compression with selective state spaces

    Shiyu Qin, Jinpeng Wang, Yiming Zhou, Bin Chen, Tianci Luo, Baoyi An, Tao Dai, Shutao Xia, and Yaowei Wang. Mambavc: Learned visual compression with selective state spaces. arXiv preprint arXiv:2405.15413, 2024. 6, 7, 13

    2024 arXiv

  35. [43]

    Skodras, C

    A. Skodras, C. Christopoulos, and T. Ebrahimi. The JPEG 2000 still image compression standard. IEEE Signal Pro- cessing Magazine, 18(5):36–58, 2001. 1

    2000

  36. [44]

    Neural video compression with context modulation

    Chuanbo Tang, Zhuoyuan Li, Yifan Bian, Li Li, and Dong Liu. Neural video compression with context modulation. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 12553–12563, 2025. 2

    2025

  37. [45]

    The JPEG still picture compression standard

    Gregory K Wallace. The JPEG still picture compression standard. Communications of the ACM, 34(4):30–44, 1991. 1

    1991

  38. [46]

    Asymllic: Asymmetric lightweight learned image compression

    Shen Wang, Zhengxue Cheng, Donghui Feng, Guo Lu, Li Song, and Wenjun Zhang. Asymllic: Asymmetric lightweight learned image compression. In 2024 IEEE Inter- national Conference on Visual Communications and Image Processing (VCIP), pages 1–5, 2024. 2

    2024

  39. [47]

    Mul- tiscale structural similarity for image quality assessment

    Zhou Wang, Eero P Simoncelli, and Alan C Bovik. Mul- tiscale structural similarity for image quality assessment. In The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003, pages 1398–1402. Ieee, 2003. 5

    2003

  40. [48]

    On uniform scalar quantization for learned image compression

    Haotian Zhang, Li Li, and Dong Liu. On uniform scalar quantization for learned image compression. arXiv preprint arXiv:2309.17051, 2023. 2

    2023 arXiv

  41. [49]

    Generalized Gaus- sian model for learned image compression

    Haotian Zhang, Li Li, and Dong Liu. Generalized Gaus- sian model for learned image compression. arXiv preprint arXiv:2411.19320, 2024. 2, 3

    2024 arXiv

  42. [50]

    Practical learned image compression with online encoder optimization

    Haotian Zhang, Feihong Mei, Junqi Liao, Li Li, Houqiang Li, and Dong Liu. Practical learned image compression with online encoder optimization. In Picture Coding Symposium (PCS), pages 1–5, 2024. 5

    2024

  43. [51]

    Learn- ing switchable priors for neural image compression

    Haotian Zhang, Yuqi Li, Li Li, and Dong Liu. Learn- ing switchable priors for neural image compression. IEEE Transactions on Circuits and Systems for Video Technology,

  44. [52]

    Transformer- based transform coding

    Yinhao Zhu, Yang Yang, and Taco Cohen. Transformer- based transform coding. In International Conference on Learning Representations (ICLR), 2022. 2

    2022

  45. [53]

    kMACs/pixel vs. BD-Rate

    Renjie Zou, Chunfeng Song, and Zhaoxiang Zhang. The devil is in the details: Window-based attention for image compression. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 17471–17480, 2022. 2, 6, 7, 12, 13 10 Learned Image Compression with Hiera...

    2022

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.