Stable Diffusion is a Natural Cross-Modal Decoder for Layered AI-generated Image Compression

Qi Mao; Ruijie Chen; Zhengxue Cheng

REVIEW 3 major objections 4 minor 26 references

Stable Diffusion is a Natural Cross-Modal Decoder for Layered AI-generated Image Compression

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper proposes a scalable layered codec that encodes AI-generated images as a text prompt, an edge or pose map, and an $8\times 8$ color map, using Stable Diffusion as the decoder, and reports reconstruction below 0.02 bpp that…

desk verdict A promising layered generative codec for AI images whose 'texture layer' is actually a coarse color prior, and whose evaluation is too thin to support the current claims. read the letter →

arxiv 2412.12982 v1 pith:L6HEWUYR submitted 2024-12-17 eess.IV cs.CV

Ruijie Chen , Qi Mao , Zhengxue Cheng This is my paper · ORCID

classification eess.IVcs.CV

keywords AI-generatedimagecompressioncross-modallayeredbitstreamStableDiffusiontext-to-imagesemanticpriorultra-lowbitrateediting

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 claims that AI-generated images can be compressed far below conventional codec bitrates by encoding them as three human-comprehensible pieces: a text caption, a structure map (edges or human pose keypoints), and an $8\times 8$ color grid. Stable Diffusion, conditioned through T2I-Adapter, then acts as the decoder, regenerating the image from these priors. On the AGIQA-3K corpus the scheme operates below 0.02 bpp and, by the paper's measurements, beats VVC and JPEG2000 in perceptual quality and roughly matches PerCo. Because the bitstream layers are human-comprehensible, the representation also supports editing operations—structure changes, texture swaps, and object erasing—without full image decoding.

What carries the argument

The machinery is the layered bitstream plus the fixed pretrained Stable Diffusion decoder. Semantic text from BLIP-2 sets content; PiDiNet edges or OpenPose keypoints supply geometry; the $8\times 8$ color map supplies coarse color; T2I-Adapter feeds the structure and texture conditions into Stable Diffusion while the prompt maintains semantics. The scalable reconstruction ladder—semantic only, semantic plus structure, semantic plus structure plus texture—carries the claim, since each added layer is encoded in a compact form and the paper reports distortion decreasing at each stage.

What would settle it

Replace the $8\times 8$ color map with a constant map holding the image's mean color while leaving the text and structure layers unchanged; if DISTS, FID, and ClipSIM barely move, the color map is not preserving texture and the texture-layer claim is falsified.

Watch

Extended reading notes

Core claim

The central discovery is that Stable Diffusion, despite being trained for generation rather than compression, behaves as a natural cross-modal decoder when its input conditions are chosen to be cheaply encodable priors. The paper shows that a semantic prior (BLIP-2 text prompt), a structural prior (PiDiNet edge map or OpenPose pose map), and a texture prior (an $8\times 8$ color map downsampled by a factor of 64) can be compressed into a scalable bitstream that reconstructs AIGIs with increasing fidelity as layers are added. The paper reports this method reaches ultra-low bitrates under 0.02 bpp while improving on VVC and JPEG2000 in perceptual and semantic metrics on the AGIQA-3K dataset, and it enables direct editing of the compressed representation without decoding.

Load-bearing premise

The texture layer rests on the assumption that downsampling an image to an $8\times 8$ grid of color values and letting Stable Diffusion fill in the rest preserves the original texture, even though that grid records only 64 average colors and cannot carry fine detail.

Editorial extensions

If this is right

On the AGIQA-3K dataset, AI-generated images can be reconstructed below 0.02 bpp with better perceptual quality than VVC and JPEG2000 and quality comparable to PerCo, according to the reported metrics.
The compressed bitstream supports structure manipulation, texture synthesis, and object erasing by editing the edge or pose map and the color grid, then rerunning the decoder without full image decoding.
Reconstruction quality is scalable: adding the structure layer and then the texture layer to the semantic prompt reduces distortion, so one bitstream serves multiple fidelity levels.
The approach establishes AI-generated image compression as a distinct problem in which decoding is a generation step conditioned on transmitted priors.

Reading between the lines

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

The paper leaves implicit that the method's fidelity ceiling is set by Stable Diffusion's internal prior: images whose content lies far outside that prior would come back plausible rather than faithful.
A decisive ablation the paper does not report would replace the $8\times 8$ color grid with a constant mean-color map while keeping the prompt and structure layers fixed; if quality metrics stay essentially unchanged, the texture layer is not carrying texture information.
The same layered pipeline could be run on natural images, and a large performance gap versus AI-generated images would confirm that the decoder is exploiting the fact that both the compressed images and the decoder come from the same generative distribution.
The editing results hint that the compressed representation itself could serve as the native editing surface in a joint codec-and-editor system, avoiding decode-edit-reencode loops.

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.

Desk Editor's Note

read the letter

Short version: this is a coherent integration paper — three human-readable priors (BLIP-2 text, PiDiNet edge or OpenPose pose, and an 8x8 color grid) decoded by Stable Diffusion with T2I-Adapter for AIGI compression — and the bitstream-level editing tricks are a nice practical addition. But the evidence is thinner than the claims, and the 'texture layer' is mischaracterized: an 8x8 grid for a 512x512 image is a coarse color prior, not a texture representation.

What's actually new: the specific three-layer scalable combination and the editing-from-bitstream operations. Prior cross-modal codecs used GAN decoders or single-modality prompts; this setup with SD as decoder is a reasonable extension. Implementation details are sufficient for reimplementation.

Soft spots. First, the texture-layer problem. Eq. (3) downsamples by 64, leaving 64 color samples. That cannot preserve local texture. The abstract says the texture layer 'preserves local textures' — unsupported. Any texture fidelity comes from Stable Diffusion's prior filling in plausible detail, not from transmitted information. The metrics (DISTS, NIQE, FID, ClipSIM) are global and cannot separate faithful reconstruction from plausible regeneration. The qualitative figures use hand-picked examples that don't stress texture-rich content.

Second, evaluation is thin. No aggregate table, no error bars, no significance tests, no released code or data. The dataset is hand-filtered to 434 of 2982 images with pose-area and category constraints, so generalization is unproven. And the text says 'significantly outperforms VVC' but the per-image DISTS values in Figure 3 include cases where VVC is better. The R-D curves are the only quantitative evidence, and they don't show per-point variance.

Third, the same color-grid issue weakens the editing story: swapping an 8x8 color map between images transfers a color grade, not texture. That's a fine feature, but it shouldn't be called texture synthesis.

What's good: the framework is coherent, the baseline comparisons are fair (PerCo is the right ultra-low-bitrate neural codec to compare against), and there's no circularity — nothing is fitted to the evaluation data. The citation pattern is fine; the one self-reference ([19]) is not load-bearing.

This paper is for people working on AIGI compression or generative codecs. It deserves a serious referee — a good reviewer could push it toward something solid — but I wouldn't cite it in the current form. Recommendation: engage, but require the authors to rename the color layer, add aggregate numbers with variance, test on a broader subset, and release code.

Referee Report

3 major / 4 minor

Summary. This manuscript proposes a layered cross-modal compression framework for AI-generated images. The encoder extracts three human-comprehensible priors: a BLIP-2 text caption (semantic layer), a PiDiNet edge map or OpenPose pose map (structure layer), and an 8x8 color map obtained by 64x downsampling (texture layer). The decoder is a pretrained Stable Diffusion 1.5 with T2I-Adapter conditioning, reconstructing images from subsets of these priors. The authors claim ultra-low bitrate compression (<0.02 bpp) with semantic and perceptual fidelity competitive with or better than VVC, JPEG2000, and PerCo, and demonstrate bitstream-level editing (structure manipulation, texture synthesis, object erasing) on qualitative examples. Experiments are on a manually filtered subset of AGIQA-3K (434 of 2982 images).

Significance. The idea is timely and the system is simple, reproducible in principle from public components, and requires no training. If the quantitative claims were substantiated, the paper would contribute a practical ultra-low-bitrate codec for AIGIs and a useful demonstration that bitstream-level editing can ride on the compressed representation. However, the evidence as presented is not yet sufficient: there are no aggregate metrics, no statistical analysis, and the texture layer is too coarse to support the claimed texture preservation. The manuscript would be strengthened by releasing the filtered dataset and code, reporting per-layer aggregate R-D data with confidence intervals, and adding texture-specific fidelity checks.

major comments (3)

[Section 2.1, Eq. (3), and Section 2.2, Eq. (6)] The 'texture layer' is an 8x8 color map formed by downsampling the image by a factor of 64; for the 512x512 inputs used here, each entry represents one average color for a 64x64 block. Such a map cannot carry local texture information, so the Abstract's claim that the texture layer 'preserves local textures' and Eq. (6)'s claim of 'high perceptual fidelity' are not supported by the bitstream construction. In the full three-layer reconstruction, any fine texture in the reconstruction is generated by Stable Diffusion's prior, not transmitted. The global metrics (DISTS, NIQE, FID, ClipSIM) cannot distinguish faithful texture reconstruction from plausible regeneration. Please add a texture-focused evaluation (e.g., local statistics or patch similarity on texture-rich crops, and an ablation replacing the color map with a fixed constant map) and discuss this limitation explicitly.
[Section 3.1 and Section 3.2, Fig. 4] The quantitative evidence for the central claim consists of R-D curves and a few qualitative examples; there is no aggregate table with mean/median bpp and metrics, no error bars, and no significance test. The dataset is reduced from 2982 to 434 images by hand-filtering with pose-area and category constraints, and no subset or code is released, so the reader cannot reproduce the reported advantage over VVC and JPEG2000. Please report full-dataset aggregate metrics (with confidence intervals) for each layer and baseline, and release the evaluation subset or explain why it cannot be released.
[Section 3.2, Fig. 4] The claimed scalability ('image fidelity improves as more structure and texture information is provided') is asserted but not quantitatively documented: the R-D curves in Fig. 4 plot only the final operating points, and the per-layer bpp/DISTS values are given only in figure captions for selected images. Provide a table with the mean bpp and mean metrics for Layer 1, Layer 2, and Layer 3, and for both pose-based and edge-based subsets, so the incremental gain of each layer can be verified.

minor comments (4)

[Throughout] The acronym 'AIGI' is sometimes written as 'AGI' (e.g., 'layered AGI compression framework' in Fig. 1); please standardize the terminology.
[Fig. 4 caption] There is a typo in the caption: 'texturte priors' should be 'texture priors'.
[References] References [11] and [23] are the same paper ('Text + Sketch: Image Compression at Ultra Low Rates'); please cite it only once.
[Section 3.3] The editing results are presented only as qualitative examples without any user study or objective editing-quality metric; consider adding at least a small quantitative comparison or noting this as a preliminary demonstration.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the framework is an empirical codec with fixed pretrained components and no fitted parameter is renamed as a prediction.

full rationale

The paper makes no formal derivation that could be circular. Each layer is an explicit encoder choice: BLIP-2 caption (Eq. 1), PiDiNet/OpenPose map (Eq. 2), and an 8x8 downsample (Eq. 3); the decoder is a fixed pretrained Stable Diffusion with T2I-Adapter and hand-set hyperparameters. No parameter is fitted to the AGIQA-3K evaluation set, and no result is obtained by substituting an equation into itself. The only co-authored prior work cited is [19], used as background for the observation that existing methods learn implicit texture features; removing it would not change any equation, bitrate, or metric, so it is not load-bearing. The skeptical concern that an 8x8 color map cannot actually convey texture is a correctness/adequacy criticism, not a circularity: it challenges whether the texture layer does what its name claims, but the claim is an empirical one and is not equivalent to the construction by definition. Comparisons to VVC, JPEG2000, and PerCo are external baselines. Hence no circular step can be quoted, and the appropriate finding is no significant circularity.

Assumptions & free parameters 9 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the reliability of several pretrained extractors and a generative decoder, plus hand-set hyperparameters. No new free parameters are learned. The main unstated modeling assumption is that an 8x8 color map plus edges or poses and a text prompt contains enough information for high-fidelity reconstruction; the paper does not ablate this or report fidelity to the original pixels.

free parameters (9)

Stable Diffusion guidance scale = 7.5
Hand-set; no ablation or sensitivity analysis reported (Section 3.1, Implementation Details).
Diffusion sampling steps = 50
Hand-set; the choice is not justified or ablated.
T2I-Adapter conditional scale = 1
Hand-set for all priors; no sensitivity analysis is given.
PiDiNet edge detection threshold = 50
Hand-set; changes the amount of structural detail sent to the decoder.
Structure map downsampling factor = 2
Edge maps are downsampled by 2 before VVC compression; this trade-off is not analyzed.
VVC quantization parameter for edge maps = 53
Hand-set QP; strongly influences the structure layer bitrate and quality.
Color map size after 64x downsampling = 8x8
512x512 images become 8x8 color grids; the choice is the core texture prior and is not ablated.
Pose keypoint coordinate precision = 2 decimal places
Quantization precision is chosen by hand before lossless compression.
Zstd compression level = 19
Used for text and pose bitstreams; a fixed implementation choice.

assumptions (6)

domain assumption Stable Diffusion 1.5 conditioned by T2I-Adapter can synthesize an image perceptually close to the original given only text, an edge or pose map, and an 8x8 color map.
This is the core enabling assumption; no fidelity bound or ablation is given.
domain assumption BLIP-2 captions preserve the semantic content needed for reconstruction.
The semantic layer relies entirely on the caption's accuracy and completeness.
domain assumption PiDiNet edge maps or OpenPose keypoint maps are sufficient geometric structure for reconstruction.
The structure layer assumes these maps capture all geometry the decoder needs.
domain assumption An 8x8 color map is an adequate texture representation for high perceptual fidelity.
The paper calls this a texture layer, but a 64-sample color grid encodes only global color, not local texture.
domain assumption The perceptual metrics DISTS, NIQE, FID, and CLIPSim measure the quality the paper claims.
All quantitative conclusions are drawn from these metrics, with no human study or pixel-fidelity check.
domain assumption The hand-selected subset of AGIQA-3K is representative of AI-generated images.
Only 434 of 2982 images are used, filtered by category and pose occupancy, so generalizability is asserted rather than demonstrated.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Stable Diffusion is a Natural Cross-Modal Decoder for Layered AI-generated Image Compression." pith.science (2026). https://pith.science/paper/L6HEWUYR

@misc{pith2026241212982,
  author       = {Pith},
  title        = {Pith review of: Stable Diffusion is a Natural Cross-Modal Decoder for Layered AI-generated Image Compression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6HEWUYR}},
  note         = {Machine review of arXiv:2412.12982}
}

read the original abstract

Recent advances in Artificial Intelligence Generated Content (AIGC) have garnered significant interest, accompanied by an increasing need to transmit and compress the vast number of AI-generated images (AIGIs). However, there is a noticeable deficiency in research focused on compression methods for AIGIs. To address this critical gap, we introduce a scalable cross-modal compression framework that incorporates multiple human-comprehensible modalities, designed to efficiently capture and relay essential visual information for AIGIs. In particular, our framework encodes images into a layered bitstream consisting of a semantic layer that delivers high-level semantic information through text prompts; a structural layer that captures spatial details using edge or skeleton maps; and a texture layer that preserves local textures via a colormap. Utilizing Stable Diffusion as the backend, the framework effectively leverages these multimodal priors for image generation, effectively functioning as a decoder when these priors are encoded. Qualitative and quantitative results show that our method proficiently restores both semantic and visual details, competing against baseline approaches at extremely low bitrates ( <0.02 bpp). Additionally, our framework facilitates downstream editing applications without requiring full decoding, thereby paving a new direction for future research in AIGI compression.

Figures

Figures reproduced from arXiv: 2412.12982 by the authors.

**Figure 2.** The qualitative comparison results of JPEG2000, VVC, Perco and our [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** The qualitative comparison results of JPEG2000, VVC, Perco and our [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: The R-D performance of JPEG2000, VVC, Perco and our method on the [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

**Figure 5.** Figure 5: Results of Structure Manipulation, Texture Synthesis and Object Erasing. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 23 canonical work pages

[19]

Disentangled visual representations for extreme human body video compression,

Ruofan Wang, Qi Mao, Shiqi Wang, Chuanmin Jia, Ronggang Wang, and Siwei Ma, “Disentangled visual representations for extreme human body video compression,” in 2022 IEEE International Conference on Multimedia and Expo (ICME) , 2022, pp. 1–6

work page 2022
[1]

High-resolution image synthesis with latent diffusion models,

Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨ orn Om- mer, “High-resolution image synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 10684–10695

work page 2022
[2]

Adding conditional control to text-to-image diffusion models,

Lvmin Zhang, Anyi Rao, and Maneesh Agrawala, “Adding conditional control to text-to-image diffusion models,” 2023

2023
[3]

T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models,

Chong Mou, Xintao Wang, Liangbin Xie, Yanze Wu, Jian Zhang, Zhongang Qi, and Ying Shan, “T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models,” in Proceedings of the AAAI Conference on Artificial Intelligence, 2024, vol. 38, pp. 4296–4304

work page 2024
[4]

Cross modal compression: Towards human-comprehensible semantic compression,

Jiguo Li, Chuanmin Jia, Xinfeng Zhang, Siwei Ma, and Wen Gao, “Cross modal compression: Towards human-comprehensible semantic compression,” in Proceedings of the 29th ACM international conference on multimedia , 2021, pp. 4230–4238

work page 2021
[5]

Cross modal compression with variable rate prompt,

Junlong Gao, Jiguo Li, Chuanmin Jia, Shanshe Wang, Siwei Ma, and Wen Gao, “Cross modal compression with variable rate prompt,” IEEE Trans. on Multimedia , 2023

work page 2023
[6]

Rate-distortion optimized cross modal compression with multiple domains,

Junlong Gao, Chuanmin Jia, Zhimeng Huang, Shanshe Wang, Siwei Ma, and Wen Gao, “Rate-distortion optimized cross modal compression with multiple domains,” IEEE Trans. on Circuits and Systems for Video Technology , 2024

work page 2024
[7]

Rethinking semantic image compression: Scalable representation with cross-modality transfer,

Pingping Zhang, Shiqi Wang, Meng Wang, Jiguo Li, Xu Wang, and Sam Kwong, “Rethinking semantic image compression: Scalable representation with cross-modality transfer,” IEEE Trans. on Circuits and Systems for Video Technology , vol. 33, no. 8, pp. 4441–4445, 2023

work page 2023

Show all 26 references

[8]

Generative adversarial networks,

Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio, “Generative adversarial networks,” Communications of the ACM , vol. 63, no. 11, pp. 139–144, 2020

2020
[9]

To- wards image compression with perfect realism at ultra-low bitrates,

Marl` ene Careil, Matthew J. Muckley, Jakob Verbeek, and St´ ephane Lathuili` ere, “To- wards image compression with perfect realism at ultra-low bitrates,” in ICLR, 2024

2024
[10]

Misc: Ultra-low bitrate image semantic compression driven by large multimodal model,

Chunyi Li, Guo Lu, Donghui Feng, Haoning Wu, Zicheng Zhang, Xiaohong Liu, Guang- tao Zhai, Weisi Lin, and Wenjun Zhang, “Misc: Ultra-low bitrate image semantic compression driven by large multimodal model,” 2024. 9

2024
[11]

Text + sketch: Image compression at ultra low rates,

Eric Lei, Yi˘ git Berkay Uslu, Hamed Hassani, and Shirin Saeedi Bidokhti, “Text + sketch: Image compression at ultra low rates,” 2023

2023
[12]

David Marr, VISION: A Computational Investigation into the Human Representation and Processing of Visual Information , MIT Press, 2010

2010
[13]

Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models,

Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi, “Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models,” in International conference on machine learning . PMLR, 2023, pp. 19730–19742

2023
[14]

Rfc 8878: Zstandard compression and the ’applica- tion/zstd’ media type,

Y. Collet and M. Kucherawy, “Rfc 8878: Zstandard compression and the ’applica- tion/zstd’ media type,” 2021

2021
[15]

Pixel difference networks for efficient edge detection,

Zhuo Su, Wenzhe Liu, Zitong Yu, Dewen Hu, Qing Liao, Qi Tian, Matti Pietik¨ ainen, and Li Liu, “Pixel difference networks for efficient edge detection,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 5117–5127

2021
[16]

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

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 applications,” IEEE Trans. on Circuits and Systems for Video Technology , vol. 31, no. 10, pp. 3736–3764, 2021

2021
[17]

Openpose: Re- altime multi-person 2d pose estimation using part affinity fields,

Z. Cao, G. Hidalgo Martinez, T. Simon, S. Wei, and Y. A. Sheikh, “Openpose: Re- altime multi-person 2d pose estimation using part affinity fields,” IEEE Trans. on Pattern Analysis and Machine Intelligence , 2019

2019
[18]

Thousand to one: Semantic prior modeling for conceptual coding,

Jianhui Chang, Zhenghui Zhao, Lingbo Yang, Chuanmin Jia, Jian Zhang, and Siwei Ma, “Thousand to one: Semantic prior modeling for conceptual coding,” in 2021 IEEE International Conference on Multimedia and Expo (ICME) , 2021, pp. 1–6

2021
[20]

Conceptual compression via deep structure and texture synthesis,

Jianhui Chang, Zhenghui Zhao, Chuanmin Jia, Shiqi Wang, Lingbo Yang, Qi Mao, Jian Zhang, and Siwei Ma, “Conceptual compression via deep structure and texture synthesis,” TIP, vol. 31, pp. 2809–2823, 2022

2022
[21]

Agiqa-3k: An open database for ai-generated im- age quality assessment,

Chunyi Li, Zicheng Zhang, Haoning Wu, Wei Sun, Xiongkuo Min, Xiaohong Liu, Guangtao Zhai, and Weisi Lin, “Agiqa-3k: An open database for ai-generated im- age quality assessment,” IEEE Trans. on Circuits and Systems for Video Technology , 2023

2023
[22]

Gans trained by a two time-scale update rule converge to a local nash equilibrium,

Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” Advances in neural information processing systems , vol. 30, 2017

2017
[23]

Text+ sketch: Image compression at ultra low rates,

Eric Lei, Yi˘ git Berkay Uslu, Hamed Hassani, and Shirin Saeedi Bidokhti, “Text+ sketch: Image compression at ultra low rates,” in ICML 2023 Workshop on Neural Compression: From Information Theory to Applications , 2023

2023
[24]

Image quality assess- ment: Unifying structure and texture similarity,

Keyan Ding, Kede Ma, Shiqi Wang, and Eero P Simoncelli, “Image quality assess- ment: Unifying structure and texture similarity,” IEEE Trans. on pattern analysis and machine intelligence , vol. 44, no. 5, pp. 2567–2581, 2020

2020
[25]

Making a “completely blind

Anish Mittal, Rajiv Soundararajan, and Alan C. Bovik, “Making a “completely blind” image quality analyzer,” IEEE Signal Processing Letters , vol. 20, no. 3, pp. 209–212, 2013

2013
[26]

The jpeg 2000 still image compression standard,

A. Skodras, C. Christopoulos, and T. Ebrahimi, “The jpeg 2000 still image compression standard,” IEEE Signal Processing Magazine , vol. 18, no. 5, pp. 36–58, 2001. 10

2000

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [19]

Disentangled visual representations for extreme human body video compression,

Ruofan Wang, Qi Mao, Shiqi Wang, Chuanmin Jia, Ronggang Wang, and Siwei Ma, “Disentangled visual representations for extreme human body video compression,” in 2022 IEEE International Conference on Multimedia and Expo (ICME) , 2022, pp. 1–6

work page 2022

[2] [1]

High-resolution image synthesis with latent diffusion models,

Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨ orn Om- mer, “High-resolution image synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 10684–10695

work page 2022

[3] [2]

Adding conditional control to text-to-image diffusion models,

Lvmin Zhang, Anyi Rao, and Maneesh Agrawala, “Adding conditional control to text-to-image diffusion models,” 2023

2023

[4] [3]

T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models,

Chong Mou, Xintao Wang, Liangbin Xie, Yanze Wu, Jian Zhang, Zhongang Qi, and Ying Shan, “T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models,” in Proceedings of the AAAI Conference on Artificial Intelligence, 2024, vol. 38, pp. 4296–4304

work page 2024

[5] [4]

Cross modal compression: Towards human-comprehensible semantic compression,

Jiguo Li, Chuanmin Jia, Xinfeng Zhang, Siwei Ma, and Wen Gao, “Cross modal compression: Towards human-comprehensible semantic compression,” in Proceedings of the 29th ACM international conference on multimedia , 2021, pp. 4230–4238

work page 2021

[6] [5]

Cross modal compression with variable rate prompt,

Junlong Gao, Jiguo Li, Chuanmin Jia, Shanshe Wang, Siwei Ma, and Wen Gao, “Cross modal compression with variable rate prompt,” IEEE Trans. on Multimedia , 2023

work page 2023

[7] [6]

Rate-distortion optimized cross modal compression with multiple domains,

Junlong Gao, Chuanmin Jia, Zhimeng Huang, Shanshe Wang, Siwei Ma, and Wen Gao, “Rate-distortion optimized cross modal compression with multiple domains,” IEEE Trans. on Circuits and Systems for Video Technology , 2024

work page 2024

[8] [7]

Rethinking semantic image compression: Scalable representation with cross-modality transfer,

Pingping Zhang, Shiqi Wang, Meng Wang, Jiguo Li, Xu Wang, and Sam Kwong, “Rethinking semantic image compression: Scalable representation with cross-modality transfer,” IEEE Trans. on Circuits and Systems for Video Technology , vol. 33, no. 8, pp. 4441–4445, 2023

work page 2023