Pith. sign in

REVIEW 3 major objections 6 minor 6 cited by

Semantics-Guided Diffusion for Deep Joint Source-Channel Coding in Wireless Image Transmission

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read SGD-JSCC transmits a text caption or edge map alongside deep joint source–channel coded image features and uses a semantics-guided diffusion model to denoise the channel output, claiming perceptual gains over prior DeepJSCC schemes at SNR…

desk verdict A solid, referee-worthy JSCC-diffusion paper whose main edge-guided results hold up; the free-text side channel fails at the low-SNR regime where text guidance is claimed to shine, but that flaw mostly hits an ablation, not the headline comparison. read the letter →

arxiv 2501.01138 v2 pith:7IDQ52UG submitted 2025-01-02 cs.IT eess.SPmath.IT

Maojun Zhang , Haotian Wu , Guangxu Zhu , Richeng Jin , Xiaoming Chen , Deniz Gündüz This is my paper · ORCID
classification cs.ITeess.SPmath.IT
keywords semantics-guideddiffusiondeepjointsource-channelcodingwirelessimagetransmissionchanneldenoisingmodelsemanticsideinformationpilot-freeCSIestimationfastfadingadaptationperceptualquality
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 aims to move deep joint source–channel coding (DeepJSCC) from transmitting image features alone to transmitting image features plus a small amount of semantic side information—a text caption or an edge map—that is used to guide a diffusion model acting as a channel denoiser. The central proposal, SGD-JSCC, treats the noisy channel output as an intermediate state of a diffusion process and iteratively removes channel noise before the JSCC decoder, rather than using diffusion to post-process a distorted reconstruction. The authors argue that this semantic conditioning turns a generative model from a source of plausible but arbitrary images into a faithful reconstructor, and that it makes the system channel-adaptive without pilots or fine-tuning. If the claim holds, wireless image transmission can preserve recognizable semantics at signal-to-noise ratios as low as −15 dB, where conventional DeepJSCC reconstructs images that lose key content. The paper reports gains over prior diffusion-based and attention-based DeepJSCC schemes on perceptual metrics (LPIPS, FID, CLIP score) at a bandwidth compression ratio of 1/20, while acknowledging that pixel-level PSNR remains below a leading baseline.

What carries the argument

The mechanism that carries the argument is the placement and conditioning of the diffusion denoiser. Instead of refining the decoder output, the denoiser operates on the equalized channel output before the JSCC decoder, viewing the channel output as one intermediate noisy state in a variance-preserving forward diffusion with a continuous sigmoid noise schedule; the denoiser predicts the clean latent from that state. Conditioning comes from semantic side information: text embeddings serve as keys and values in cross-attention layers, while edge maps enter through a frozen copy of the base network plus a zero-initialized linear fusion branch. The same diffusion schedule is used for channel adaptation: a learned estimator maps the received signal to the equivalent SNR and phase, the inverse schedule maps SNR to a continuous starting timestep, and the fast-fading extension equalizes per-symbol noise levels by adding noise to over-clean elements, then updates only elements whose noise level is above the next target level.

What would settle it

Run SGD-JSCC at SNR = −15 dB while transmitting the text caption through the same fading channel, or at a finite rate, rather than assuming perfect delivery, and compare the text-guided variant against the edge-map-only and unconditional variants on LPIPS, FID, and CLIP score; if the text-guided gains disappear or reverse, the semantics-guidance claim at extreme low SNR collapses.

Watch

Extended reading notes

Core claim

SGD-JSCC's core discovery is that the channel noise can be matched to the forward noise of a diffusion model and removed under semantic guidance, so that the JSCC encoder and decoder can be trained once for one channel state while the diffusion denoiser absorbs all channel variation. Two semantic modalities are supported: coarse text descriptions, extracted by an image-captioning model and injected through cross-attention, and fine edge maps, transmitted through a dedicated JSCC encoder and injected through a zero-initialized structural-control branch. In slow fading, the receiver estimates the instantaneous SNR and phase directly from the normalized channel output, converts the estimated SNR to a continuous noise level via the inverse scheduling function, and starts denoising at the matched timestep, eliminating pilot-based channel estimation. In fast fading, a water-filling-inspired procedure adds carefully chosen Gaussian noise to cleaner symbols to equalize their noise levels to the diffusion schedule, so a denoiser trained under AWGN can be applied without retraining. The reported result is that this scheme outperforms ADJSCC, JSCCformer, DeepJSCC-Diff, JSCCDiff, and a non-diffusion VAE variant on LPIPS, FID, and CLIP score over the tested conditions, with visual examples showing semantic content such as a bear in grass preserved at −10 dB where unconditional denoising reconstructs an unidentified animal.

Load-bearing premise

The load-bearing premise is that the text description reaches the receiver perfectly and costs no channel resources; if that side channel is noisy or rate-limited, the text-guided gains reported at very low SNR may not be achievable in practice.

Editorial extensions

If this is right

  • In slow fading, the system can operate without pilot transmissions: SNR and phase are estimated from the received latent, and the paper reports these estimates are accurate enough that reconstruction quality is nearly the same with and without receiver SNR knowledge.
  • The JSCC encoder–decoder is channel-agnostic: once trained under a fixed AWGN SNR, it can be reused across channel conditions and semantic metrics, with only the diffusion denoiser and guidance modules tailored.
  • In fast fading, a diffusion denoiser trained for AWGN can be applied without fine-tuning via the water-filling update rule, and the paper reports only a slight performance decline as fading block length increases from 1 to 1024.
  • Masking parts of the latent feature during diffusion training enables rate-adaptive transmission, where dropping the lowest-L2-norm tokens degrades quality less than random dropping, suggesting an importance-based resource allocation rule.
  • The semantic side information is what preserves meaning at extreme low SNR: at −15 dB the scheme retains recognizable objects, whereas conventional DeepJSCC baselines lose semantics even when their PSNR is higher.

Reading between the lines

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

  • The paper's text-based variant assumes the caption is received perfectly and costs nothing; transmitting the caption through the same channel at finite rate is a testable extension that would reveal whether the reported low-SNR gains survive when the guidance itself is noisy.
  • Because the diffusion prior is trained on open-domain image–text data, the receiver can generate content that is plausible but absent from the source; a natural next evaluation is to measure object-level consistency, not just global perceptual scores, between the reconstruction and the original.
  • The continuous noise-level matching and water-filling update are channel-agnostic in principle; the paper targets point-to-point image links, but the same mechanism is a plausible starting point for MIMO, OFDM, or non-image modalities, which the paper does not test.
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 SGD-JSCC, a deep joint source-channel coding scheme in which a diffusion model is used to denoise the JSCC channel output before decoding, guided by semantic side information in the form of either a text description or an edge map. For slow fading channels the scheme estimates the instantaneous SNR and phase directly from the equalized channel output, so that the diffusion starting step is matched continuously to the channel state without pilot overhead; for fast fading channels it introduces a water-filling-inspired denoising procedure that equalizes per-symbol noise levels so that a DM trained under AWGN can be applied without fine-tuning. The numerical section compares SGD-JSCC with ADJSCC, JSCCformer, DeepJSCC-Diff, JSCCDiff, and VAEJSCC on COCO2017 and Kodak, reporting gains in perceptual metrics (LPIPS, FID, CLIP) at low SNR, and includes ablations of text versus edge-map guidance, blind CSI estimation accuracy, masking strategies, and fast-fading behavior.

Significance. If the reported results hold, the paper makes a useful step toward making diffusion-based JSCC practical: the continuous step-matching formulation (Eq. (20)) is a clean, parameter-free mapping from SNR to diffusion noise level, and the water-filling extension in Section VI is an interesting training-free way to reuse a slow-fading DM in fast-fading channels. The manuscript also contains extensive comparisons and ablations, and the authors promise public code and checkpoints, which would facilitate reproducibility. The main limitation is that the text-guidance variant assumes an ideal, cost-free side channel at exactly the low-SNR regime where the paper claims its largest perceptual gains; this assumption is capacity-infeasible at SNR = -15 dB and needs to be reframed as an upper bound or accompanied by an evaluated transmission scheme. The edge-map variant is within the rate budget but its reliability at very low SNR is not directly measured.

major comments (3)
  1. [Section IV-A, footnote 2] The perfect, cost-free text side channel is capacity-infeasible in the regime the paper highlights. For a 128x128x3 image at CBR = 1/20, the system has roughly 2,458 complex channel symbols in total. At SNR = -15 dB, the Shannon capacity is approximately 0.045 bits per complex dimension, giving about 110 bits per image for the whole transmission. A typical 50-character BLIP2 caption (~400 bits) would require on the order of 8,900 channel uses, more than three times the entire CBR budget. Since Fig. 15 attributes the best low-SNR CLIP/FID performance to text guidance, those specific gains are not deployable at -15 dB. The footnote's appeal to 'specialized DeepJSCC techniques' is not backed by a scheme, simulation, or capacity analysis. Please either count the text overhead in the CBR and evaluate with corrupted text reception, or present the text-guided results explicitly as an upper bound and make the edge-map-guided variant the primary deployment claim.
  2. [Section VII-C1 and Fig. 15] The reliability of the decoded edge map at very low SNR is not measured, even though edge-map guidance is the variant used in the main -15 dB comparisons (footnote 6). The only edge-map reconstruction example, Fig. 16, is shown at SNR = -10 dB, and the paper itself states in Section VII-C1 that edge-map transmission errors can mislead the diffusion model and degrade performance at high SNR. Without a quantitative evaluation of decoded edge-map fidelity at -15 dB (e.g., BCE/Dice on the COCO validation set) or an ablation that compares ground-truth edge guidance with decoded edge guidance, the claim that semantic guidance is responsible for the -15 dB perceptual gains is not fully supported. Please add such measurements or soften the corresponding claims.
  3. [Section VII, Performance Metrics and Fig. 15] The use of CLIP score as a reconstruction metric is biased in favor of the text-guided variant because the receiver has access to the ground-truth caption. A text-conditioned diffusion model can obtain a high CLIP score by producing any image that is semantically consistent with the caption, including hallucinated content that the paper itself acknowledges can occur (Section VII-C1). The FID metric has a related realism bias. To support the perceptual-quality claims fairly, please report reconstruction-oriented metrics (LPIPS with the decoded edge map, or a faithfulness metric that penalizes caption-consistent but content-incorrect reconstructions) alongside CLIP/FID, and discuss the bias explicitly.
minor comments (6)
  1. [Section II-C and Section VII-A] The benchmark naming is inconsistent: the text alternates between 'DeepJSCC-Diff', 'JSCCDiff', and 'DiffJSCC' for what appear to be the same or closely related baselines; please align the names in the text, figures, and table captions.
  2. [Section VII-A] The sentence 'The ADJSCC scheme refers to the DeepJSCC architecture in [14]' appears to cite the wrong reference, since ADJSCC is introduced in [16]; please correct the citation.
  3. [Section IV-B and Fig. 2] The text says the edge-map JSCC model uses a ViT architecture, while Fig. 2 is labeled 'Swin Transformer'; please clarify which architecture is actually used.
  4. [Eq. (16)] As typeset, the noise coefficient in Eq. (16) appears to contain an extra factor; please verify that it matches the standard variance-preserving conditional distribution q(ft|fs).
  5. [Table I] Table I mixes layer architecture with training hyperparameters (batch size, epochs, optimizer); please separate the network specification from the training configuration.
  6. [Abstract and Section VI] The abstract and introduction emphasize 'no additional pilot transmissions for channel estimation', but the fast-fading extension in Section VI assumes perfect CSI at the receiver; please state this assumption explicitly in the contributions and conclusions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the step-matching rule is a parameter-free mapping from the channel model to the diffusion noise schedule, and the trained modules are not reused as the reported performance targets.

full rationale

The paper's claimed derivation chain is self-contained rather than circular. The central step-matching rule in Eq. (20) maps the equalized channel output's signal level alpha = |h|^2/(|h|^2 + sigma^2) to a diffusion noise level beta = sigma^2/(sigma^2 + |h|^2) by direct comparison of the channel conditional distribution in Eq. (9) with the variance-preserving forward diffusion process in Eq. (15); the sigmoid schedule S(t) is a separately chosen hyperparameterized function, and the diffusion model is trained (Algorithm 1) to predict the clean latent f0 from noisy features, not to optimize the reported PSNR, LPIPS, FID, or CLIP targets. The SNR and phase estimators are trained auxiliary modules whose outputs feed the step matcher, but their outputs are not renamed as reconstruction predictions; their accuracy and sensitivity are separately measured in Fig. 17. Semantic side information (BLIP2 text or MuGE edge map) is an input to the conditional diffusion model, and the main benchmark comparisons use the edge-map variant with its transmission cost counted inside the CBR budget, so no fitted value of the target metric is used as a prediction. The only self-citation that could be noted, reference [51] to the authors' own arXiv version, is used for 'more results on SNR estimation errors' and is not load-bearing. Footnote 2's assumption that text is transmitted perfectly is a deployability limitation at very low SNR, not a circular step, because the text-guided results are presented as an ablation and the main low-SNR claims use edge-map guidance. No equation or fitted parameter reduces the reported results to the inputs by construction.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The core algorithm is a deep learning pipeline; no new physical entities are introduced. The main free parameters are hand-chosen diffusion and guidance hyperparameters plus a reported GAN loss weight. The most fragile assumptions are the perfect-text side channel and the exact match between the equalized channel model and the diffusion forward process.

free parameters (4)
  • Sigmoid noise schedule shape (e, g, tau) = e=3, g=0, tau=0.7
    These shape parameters define the diffusion noise schedule S(t) in Eq. (18) and control the mapping from SNR to timestep used for step matching; they are chosen by hand, not derived.
  • Classifier-free guidance scalar = 4.5
    Inference-time guidance strength in the DM (Table II), tuned by the authors, affects the balance between realism and fidelity.
  • Guidance scalar = 0.3
    Additional guidance weight in Table II for semantic conditioning; hand-set.
  • GAN loss weight lambda = not reported
    Weight of the discriminator loss in Eq. (14); the value is not stated, so the exact JSCC training objective is not fully reproducible from the paper.
assumptions (5)
  • domain assumption The wireless channel after equalization is exactly a variance-preserving Gaussian corruption of the JSCC latent feature, matching the DM forward process (Eq. 9 equals Eq. 15).
    Used in Section V-B to justify step matching m = S^{-1}(1/(1+gamma)); requires perfect synchronization, known noise distribution, and correct power normalization.
  • domain assumption The DM trained on a fixed JSCC encoder at SNR=10 dB generalizes to other SNRs and to the COCO test distribution.
    The JSCC model is fixed after training on ImageNet at 10 dB (Section V-A); the DM is trained on that encoder's latents and applied to COCO at SNRs from -15 to 15 dB.
  • domain assumption Text descriptions can be transmitted to the receiver perfectly at negligible cost.
    Stated in Section IV-A footnote 2; this makes the text-guided variant free of a side-channel cost and is not valid at very low SNR without additional mechanisms.
  • domain assumption Edge map side information can be transmitted with sufficient fidelity within 1/24 of the total CBR.
    The edge map JSCC model is trained with BCE plus Dice loss (Section IV-B) and its errors are fed into the DM as guidance; if edge map errors are large, the guidance can mislead denoising.
  • standard math Diffusion theory: DDIM-type deterministic reverse sampling with sigma=0 reconstructs the clean latent (Eq. 17 with sigma_{s,t}=0).
    Adopted from Song et al. [39] and used in Eqs. (19) and (27); this is standard DDIM theory.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semantics-Guided Diffusion for Deep Joint Source-Channel Coding in Wireless Image Transmission." pith.science (2026). https://pith.science/paper/7IDQ52UG

@misc{pith2026250101138,
  author       = {Pith},
  title        = {Pith review of: Semantics-Guided Diffusion for Deep Joint Source-Channel Coding in Wireless Image Transmission},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7IDQ52UG}},
  note         = {Machine review of arXiv:2501.01138}
}
read the original abstract

Joint source-channel coding (JSCC) offers a promising avenue for enhancing transmission efficiency by jointly incorporating source and channel statistics into the system design. A key advancement in this area is the deep joint source and channel coding (DeepJSCC) technique that designs a direct mapping of input signals to channel symbols parameterized by a neural network, which can be trained for arbitrary channel models and semantic quality metrics. This paper advances the DeepJSCC framework toward a semantics-aligned, high-fidelity transmission approach, called semantics-guided diffusion DeepJSCC (SGD-JSCC). Existing schemes that integrate diffusion models (DMs) with JSCC face challenges in transforming random generation into accurate reconstruction and adapting to varying channel conditions. SGD-JSCC incorporates two key innovations: (1) utilizing some inherent information that contributes to the semantics of an image, such as text description or edge map, to guide the diffusion denoising process; and (2) enabling seamless adaptability to varying channel conditions with the help of a semantics-guided DM for channel denoising. The DM is guided by diverse semantic information and integrates seamlessly with DeepJSCC. In a slow fading channel, SGD-JSCC dynamically adapts to the instantaneous signal-to-noise ratio (SNR) directly estimated from the channel output, thereby eliminating the need for additional pilot transmissions for channel estimation. In a fast fading channel, we introduce a training-free denoising strategy, allowing SGD-JSCC to effectively adjust to fluctuations in channel gains. Numerical results demonstrate that, guided by semantic information and leveraging the powerful DM, our method outperforms existing DeepJSCC schemes, delivering satisfactory reconstruction performance even at extremely poor channel conditions.

Figures

Figures reproduced from arXiv: 2501.01138 by the authors.

Figure 1
Figure 1. Different DeepJSCC transmission paradigms. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. DeepJSCC transmission for the image edge map. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Network architecture of diffusion transformer (DiT) model. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (15 more)
Figure 5
Figure 5. Figure 5: Diffusion Denoising Procedure. Remark 1. Using diffusion for preprocessing the channel output, rather than post-processing the DeepJSCC output as in [28], [29], offers some new advantages. While DM for post-processing can leverage the prior knowledge from the learned d…
Figure 6
Figure 6. Figure 6: Block diagram of the proposed semantics-guided diffusion framework. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Comparison of step matching method [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Adaptation to the fast fading channel. where di = σ 2 |hc,i| 2+σ2 , and hc,i =  hi , i ≤ N/2 hi−N/2, else . As illustrated in [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Average reconstruction performance over AWGN channels with receiver-side SNR information. 15 10 5 0 5 10 15 SNR 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 P S N R VAEJSCC ADJSCC DiffJSCC DeepJSCC-Diff JSCCformer SGD-JSCC(proposed) (a) PSNR, AWGN w/o SNR 15 10 5 0 5 1…
Figure 10
Figure 10. Figure 10: Average reconstruction performance over AWGN channels without SNR information. 15 10 5 0 5 10 15 SNR 5 10 15 20 25 30 P S N R VAEJSCC ADJSCC DiffJSCC DeepJSCC-Diff JSCCformer SGD-JSCC(proposed) (a) PSNR, Fading w/o CSI 15 10 5 0 5 10 15 SNR 0.0 0.1 0.2 0.3 0.4 0.5 0.6…
Figure 11
Figure 11. Figure 11: Average reconstruction performance over slow Rayleigh fading channels without CSI information. σ 2 unknown.), and 3) Rayleigh fading channels without CSI (h ∈ CN (0, 1), h and σ 2 unknown.). 1) AWGN channels with receiver-side SNR information: We first evaluate the pe…
Figure 12
Figure 12. Figure 12: Examples of reconstructed images under AWGN channel with SNR [PITH_FULL_IMAGE:figures/full_fig_p012_12.png]
Figure 13
Figure 13. Figure 13: Examples of reconstructed images under AWGN channel with SNR [PITH_FULL_IMAGE:figures/full_fig_p012_13.png]
Figure 14
Figure 14. Figure 14: Examples of reconstructed images under AWGN channel with SNR [PITH_FULL_IMAGE:figures/full_fig_p012_14.png]
Figure 15
Figure 15. Figure 15: Ablation study of different guidance schemes. Original Edge Map Rec. Edge Map Unconditional Text Guided Edge Guided Text and Edge Guided [PITH_FULL_IMAGE:figures/full_fig_p013_15.png]
Figure 16
Figure 16. Figure 16: Examples of reconstructed images under different guidance methods with SNR [PITH_FULL_IMAGE:figures/full_fig_p013_16.png]
Figure 17
Figure 17. Figure 17: Performance Evaluation of Channel Estimation and Sensitivity Analysis. [PITH_FULL_IMAGE:figures/full_fig_p014_17.png]
Figure 18
Figure 18. Figure 18: Examples of reconstructed images under different mask strategies with SNR [PITH_FULL_IMAGE:figures/full_fig_p014_18.png]
Figure 19
Figure 19. Figure 19: Performance evaluation with mask operation. [PITH_FULL_IMAGE:figures/full_fig_p015_19.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

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

  1. Rethinking Multi-User Communication in Semantic Domain: Enhanced OMDMA by Shuffle-Based Orthogonalization and Diffusion Denoising

    cs.IT 2025-07 conditional novelty 6.0 of 10

    Randomly permuting each user's JSCC features turns multi-user interference into roughly Gaussian noise, letting a single pretrained encoder and diffusion denoiser serve all users.

  2. Token-Domain Multiple Access: Exploiting Semantic Orthogonality for Collision Mitigation

    cs.IT 2025-02 conditional novelty 6.0 of 10

    ToDMA lets uncoordinated devices share a token codebook and transmit non-orthogonally, then uses a pretrained transformer to repair token collisions from context.

  3. Mixture of Semantics Transmission for Generative AI-Enabled Semantic Communication Systems

    cs.IT 2025-09 conditional novelty 5.0 of 10

    A semantic communication system that transmits image ROIs at high fidelity and transmits the background as text, reconstructing the image with a diffusion model.

  4. ToDMA: Large Model-Driven Massive Token Communications for Semantic Multiple Access

    cs.IT 2025-05 conditional novelty 5.0 of 10

    ToDMA lets many devices share one wireless channel by transmitting token indices from a common codebook, recovering collisions with compressed sensing and masked-token prediction from pretrained models.

  5. Visual Fidelity Index for Generative Semantic Communications with Critical Information Embedding

    eess.IV 2025-05 conditional novelty 5.0 of 10

    The paper introduces GVIF, a mutual-information-based metric for visual fidelity in generative semantic communication, and uses it to adaptively select critical image features and compression levels for wireless transmission.

  6. Censored Sampling for Topology Design: Guiding Diffusion with Human Preferences

    cs.LG 2025-08 unverdicted novelty 4.0 of 10

    Guiding a pretrained topology-diffusion generator with human-preference reward classifiers is claimed to suppress floating-material and boundary-violation failure modes without retraining the generator.

Reference graph

Works this paper leans on

51 extracted references · 32 canonical work pages · cited by 6 Pith papers

  1. [1]

    Beyond transmitting bits: Context, semantics, and task-oriented communications,

    D. G ¨und¨uz, Z. Qin, I. E. Aguerri, H. S. Dhillon, Z. Yang, A. Yener, K. K. Wong, and C.-B. Chae, “Beyond transmitting bits: Context, semantics, and task-oriented communications,” IEEE Journal on Selected Areas in Communications, vol. 41, no. 1, pp. 5–41, 2023

    work page 2023

  2. [2]

    Deep joint source- channel coding for wireless image transmission,

    E. Bourtsoulatze, D. B. Kurka, and D. G ¨und¨uz, “Deep joint source- channel coding for wireless image transmission,” IEEE Trans. Cogn. Commun. Netw., vol. 5, no. 3, pp. 567–579, 2019

    2019

  3. [3]

    Joint source–channel coding: Fundamentals and recent progress in practical designs,

    D. G ¨und¨uz, M. A. Wigger, T.-Y . Tung, P. Zhang, and Y . Xiao, “Joint source–channel coding: Fundamentals and recent progress in practical designs,” Proceedings of the IEEE , 2024

    2024

  4. [4]

    Nonlinear transform source-channel coding for semantic communications,

    J. Dai, S. Wang, K. Tan, Z. Si, X. Qin, K. Niu, and P. Zhang, “Nonlinear transform source-channel coding for semantic communications,” IEEE J. Sel. Areas Commun. , vol. 40, no. 8, pp. 2300–2316, 2022

    2022

  5. [5]

    Transformer-aided wireless image transmission with channel feedback,

    H. Wu, Y . Shao, E. Ozfatura, K. Mikolajczyk, and D. G ¨und¨uz, “Transformer-aided wireless image transmission with channel feedback,” IEEE Trans. Wireless Commun. , early access, 2024

    work page 2024

  6. [6]

    Deep joint source-channel coding for adaptive image transmission over MIMO channels,

    H. Wu, Y . Shao, C. Bian, K. Mikolajczyk, and D. G ¨und¨uz, “Deep joint source-channel coding for adaptive image transmission over MIMO channels,” IEEE Trans. Wireless Commun. , 2024

    work page 2024

  7. [7]

    Deep joint source channel coding for wireless image transmission with OFDM,

    M. Yang, C. Bian, and H.-S. Kim, “Deep joint source channel coding for wireless image transmission with OFDM,” in Proc. IEEE International Conference on Communications(ICC) , pp. 1–6, 2021

    work page 2021

  8. [8]

    Channel-adaptive wireless image transmission with OFDM,

    H. Wu, Y . Shao, K. Mikolajczyk, and D. G ¨und¨uz, “Channel-adaptive wireless image transmission with OFDM,” IEEE Wireless Commun. Lett., vol. 11, no. 11, pp. 2400–2404, 2022

    work page 2022

Show all 51 references
  1. [9]

    Process- and-forward: Deep joint source-channel coding over cooperative relay networks,

    C. Bian, Y . Shao, H. Wu, E. Ozfatura, and D. Gunduz, “Process- and-forward: Deep joint source-channel coding over cooperative relay networks,” [Online]. Available: https://arxiv.org/abs/2403.10613, 2024

    2024 arXiv

  2. [10]

    Distributed deep joint source-channel coding over a multiple access channel,

    S. F. Yilmaz, C. Karamanlı, and D. G ¨und¨uz, “Distributed deep joint source-channel coding over a multiple access channel,” in Prof. IEEE Int’l Conf. on Comms. (ICC) , pp. 1400–1405, 2023

    2023

  3. [11]

    Model division multiple access for semantic communications,

    P. Zhang, X. Xu, C. Dong, K. Niu, H. Liang, Z. Liang, X. Qin, M. Sun, H. Chen, N. Ma, et al. , “Model division multiple access for semantic communications,” Frontiers of Information Technology & Electronic Engineering, vol. 24, no. 6, pp. 801–812, 2023

    2023

  4. [12]

    Deepjscc- q: Constellation constrained deep joint source-channel coding,

    T.-Y . Tung, D. B. Kurka, M. Jankowski, and D. G ¨und¨uz, “Deepjscc- q: Constellation constrained deep joint source-channel coding,” IEEE J. Sel. Areas Inf. Theory , vol. 3, no. 4, pp. 720–731, 2022

    2022

  5. [13]

    Joint coding-modulation for digital semantic communications via variational autoencoder,

    Y . Bo, Y . Duan, S. Shao, and M. Tao, “Joint coding-modulation for digital semantic communications via variational autoencoder,” IEEE Transactions on Communications, vol. 72, no. 9, pp. 5626–5640, 2024

    2024

  6. [14]

    Generative joint source-channel coding for semantic image transmission,

    E. Erdemir, T.-Y . Tung, P. L. Dragotti, and D. G¨und¨uz, “Generative joint source-channel coding for semantic image transmission,” IEEE J. Sel. Areas Commun., vol. 41, no. 8, pp. 2645–2657, 2023

    2023

  7. [15]

    CDDM: Channel denoising diffusion models for wireless semantic communications,

    T. Wu, Z. Chen, D. He, L. Qian, Y . Xu, M. Tao, and W. Zhang, “CDDM: Channel denoising diffusion models for wireless semantic communications,” IEEE Trans. Wireless Commun. , 2024

    2024

  8. [16]

    Wireless image transmission using deep source channel coding with attention modules,

    J. Xu, B. Ai, W. Chen, A. Yang, P. Sun, and M. Rodrigues, “Wireless image transmission using deep source channel coding with attention modules,” IEEE Trans. Circuits Syst. Video Technol. , vol. 32, no. 4, pp. 2315–2328, 2021

    2021

  9. [17]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Proc. Adv. in Neural Inf. Proc. Sys. (NeurIPS) , pp. 6840–6851, 2020

    2020

  10. [18]

    Diffusion models beat gans on image synthesis,

    P. Dhariwal and A. Nichol, “Diffusion models beat gans on image synthesis,” Proc. Advances in Neural Information Processing Systems (NeurIPS), vol. 34, pp. 8780–8794, 2021

    2021

  11. [19]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” in IEEE/CVF Conf. Comp. Vision and Pattern Recog. (CVPR), pp. 10684–10695, 2022

    2022

  12. [20]

    Pixart-δ: Fast and controllable image generation with latent consistency models,

    J. Chen, Y . Wu, S. Luo, E. Xie, S. Paul, P. Luo, H. Zhao, and Z. Li, “Pixart-δ: Fast and controllable image generation with latent consistency models,” [Online]. Available: https://arxiv.org/abs/2401.05252, 2024

    2024 arXiv

  13. [21]

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

    L. Zhang, A. Rao, and M. Agrawala, “Adding conditional control to text-to-image diffusion models,” in Proc. IEEE/CVF International Conference on Computer Vision (ICCV) , pp. 3836–3847, 2023

    2023

  14. [22]

    Text+ sketch: Image compression at ultra low rates,

    E. Lei, Y . B. Uslu, H. Hassani, and S. S. Bidokhti, “Text+ sketch: Image compression at ultra low rates,” [Online]. Available: https://arxiv.org/abs/2307.01944, 2023

    2023 arXiv

  15. [23]

    Latency-aware generative semantic communications with pre-trained diffusion models,

    L. Qiao, M. Mashhadi, Z. Gao, C. H. Foh, P. Xiao, and M. Bennis, “Latency-aware generative semantic communications with pre-trained diffusion models,” [Online]: https://arxiv.org/abs/2403.17256, 2024

    2024 arXiv

  16. [24]

    Fast-gsc: Fast and adaptive semantic transmis- sion for generative semantic communication,

    Y . Wang, W. Yang, Z. Xiong, Y . Zhao, S. Mao, T. Q. Quek, and H. V . Poor, “Fast-gsc: Fast and adaptive semantic transmis- sion for generative semantic communication,” [Online]. Available: https://arxiv.org/abs/2407.15395, 2024

    2024

  17. [25]

    Diff-go: Diffusion goal-oriented communications to achieve ultra-high spectrum efficiency,

    A. Wijesinghe, S. Zhang, S. Wanninayaka, W. Wang, and Z. Ding, “Diff-go: Diffusion goal-oriented communications to achieve ultra-high spectrum efficiency,” [Online]: https://arxiv.org/abs/2312.02984, 2023

    2023 arXiv

  18. [26]

    Lossy image compression with conditional diffusion models,

    R. Yang and S. Mandt, “Lossy image compression with conditional diffusion models,” Proc. Advances in Neural Information Processing Systems (NeurIPS), Vancouver, Canada, 2024

    2024

  19. [27]

    A hybrid wireless image transmission scheme with diffusion,

    X. Niu, X. Wang, D. G ¨und¨uz, B. Bai, W. Chen, and G. Zhou, “A hybrid wireless image transmission scheme with diffusion,” in IEEE Int’l Wrks. on Sig. Proc. Adv. in Wireless Comms. (SPAWC) , pp. 86–90, 2023

    2023

  20. [28]

    High perceptual quality wireless image delivery with denoising diffusion models,

    S. F. Yilmaz, X. Niu, B. Bai, W. Han, L. Deng, and D. Gunduz, “High perceptual quality wireless image delivery with denoising diffusion models,” [Online]. Available: https://arxiv.org/abs/2309.15889, 2023

    2023 arXiv

  21. [29]

    Commin: Semantic image communications as an inverse problem with inn-guided diffusion models,

    J. Chen, D. You, D. G ¨und¨uz, and P. L. Dragotti, “Commin: Semantic image communications as an inverse problem with inn-guided diffusion models,” in IEEE Int’l Conf. on Acous., Speech and Sig. Proc. (ICASSP), pp. 6675–6679, Seoul, Korea, 2024

    2024

  22. [30]

    Diffusion-aided joint source channel coding for high realism wireless image transmission,

    M. Yang, B. Liu, B. Wang, and H.-S. Kim, “Diffusion-aided joint source channel coding for high realism wireless image transmission,” arXiv preprint arXiv:2404.17736, 2024

    2024 arXiv

  23. [31]

    Latent diffusion model-enabled real-time semantic communication consider- ing semantic ambiguities and channel noises,

    J. Pei, F. Cheng, P. Wang, H. Tabassum, and D. Shi, “Latent diffusion model-enabled real-time semantic communication consider- ing semantic ambiguities and channel noises,” [Online]. Available: https://arxiv.org/abs/2406.06644, 2024

    2024 arXiv

  24. [32]

    Hybrid digital-analog semantic communications,

    H. Xie, Z. Qin, Z. Han, and K. B. Letaief, “Hybrid digital-analog semantic communications,” arXiv preprint arXiv:2405.12580 , 2024

    2024 arXiv

  25. [33]

    Diffusion- driven semantic communication for generative models with bandwidth constraints,

    L. Guo, W. Chen, Y . Sun, B. Ai, N. Pappas, and T. Quek, “Diffusion- driven semantic communication for generative models with bandwidth constraints,” [Online], https://arxiv.org/abs/2407.18468, 2024

    2024 arXiv

  26. [34]

    Diffcom: Channel received signal is a natural condition to guide diffusion posterior sampling,

    S. Wang, J. Dai, K. Tan, X. Qin, K. Niu, and P. Zhang, “Diffcom: Channel received signal is a natural condition to guide diffusion posterior sampling,” [Online]. Available: https://arxiv.org/abs/2406.07390, 2024

    2024 arXiv

  27. [35]

    Extreme generative image compression by learning text embedding from diffusion models,

    Z. Pan, X. Zhou, and H. Tian, “Extreme generative image compression by learning text embedding from diffusion models,” [Online]. Available: https://arxiv.org/abs/2211.07793, 2022

    2022 arXiv

  28. [36]

    Towards image compression with perfect realism at ultra-low bitrates,

    M. Careil, M. J. Muckley, J. Verbeek, and S. Lathuili `ere, “Towards image compression with perfect realism at ultra-low bitrates,” in in Proc. International Conference on Learning Representations (ICLR) , 2023

    2023

  29. [37]

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

    J. Li, D. Li, S. Savarese, and S. Hoi, “Blip-2: Bootstrapping language- image pre-training with frozen image encoders and large language models,” in Proc. International conference on machine learning(ICML) , pp. 19730–19742, Honolulu, USA, 2023

    2023

  30. [38]

    Muge: Multiple granularity edge detection,

    C. Zhou, Y . Huang, M. Pu, Q. Guan, R. Deng, and H. Ling, “Muge: Multiple granularity edge detection,” in IEEE/CVF Conf. on Computer Vision and Pattern Recog. (CVPR) , pp. 25952–25962, 2024

    2024

  31. [39]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” [Online]. Available: https://arxiv.org/abs/2010.02502, 2020

    2010 arXiv

  32. [40]

    On the importance of noise scheduling for diffusion models,

    T. Chen, “On the importance of noise scheduling for diffusion models,” [Online]. Available: https://arxiv.org/abs/2301.10972, 2023

    2023 arXiv

  33. [41]

    Scalable diffusion models with transformers,

    W. Peebles and S. Xie, “Scalable diffusion models with transformers,” in Proc. IEEE/CVF International Conference on Computer Vision (CVPR), pp. 4195–4205, 2023

    2023

  34. [42]

    Pixart- α: Fast training of diffusion trans- former for photorealistic text-to-image synthesis,

    J. Chen, J. Yu, C. Ge, L. Yao, E. Xie, Y . Wu, Z. Wang, J. Kwok, P. Luo, H. Lu, et al. , “Pixart- α: Fast training of diffusion trans- former for photorealistic text-to-image synthesis,” [Online]. Available: https://arxiv.org/abs/2310.00426, 2023

    arXiv

  35. [43]

    MDTv2: Masked Diffu- sion Transformer is a Strong Image Synthesizer,

    S. Gao, P. Zhou, M.-M. Cheng, and S. Yan, “MDTv2: Masked Diffu- sion Transformer is a Strong Image Synthesizer,” [Online]. Available: https://arxiv.org/abs/2303.14389, 2023

    2023 arXiv

  36. [44]

    An snr estimation technique based on deep learning,

    K. Yang, Z. Huang, X. Wang, and F. Wang, “An snr estimation technique based on deep learning,” Electronics, vol. 8, no. 10, p. 1139, 2019

    2019

  37. [45]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo,et al., “Segment anything,” in in Proc. IEEE/CVF International Conference on Computer Vision (CVPR), pp. 4015–4026, 2023

    2023

  38. [46]

    Journeydb: A benchmark for generative image understanding,

    K. Sun, J. Pan, Y . Ge, H. Li, H. Duan, X. Wu, R. Zhang, A. Zhou, Z. Qin, Y . Wang, et al. , “Journeydb: A benchmark for generative image understanding,” Proc. Advances in Neural Information Processing Systems (NeurIPS), vol. 36, Vancouver, Canada, 2024

    2024

  39. [47]

    Conceptual 12m: Pushing web-scale image-text pre-training to recognize long-tail visual concepts,

    S. Changpinyo, P. Sharma, N. Ding, and R. Soricut, “Conceptual 12m: Pushing web-scale image-text pre-training to recognize long-tail visual concepts,” in IEEE/CVF Conf. on Computer Vision and Pattern Recog. (CVPR), pp. 3558–3568, 2021

    2021

  40. [48]

    Datacomp: In search of the next generation of multimodal datasets,

    S. Y . Gadre, G. Ilharco, A. Fang, J. Hayase, G. Smyrnis, T. Nguyen, R. Marten, M. Wortsman, D. Ghosh, J. Zhang, et al. , “Datacomp: In search of the next generation of multimodal datasets,” Advances in Neural Inf. Proc. Systems (NeurIPS) , vol. 36, 2024

    2024

  41. [49]

    Celebv-hq: A large-scale video facial attributes dataset,

    H. Zhu, W. Wu, W. Zhu, L. Jiang, S. Tang, L. Zhang, Z. Liu, and C. C. Loy, “Celebv-hq: A large-scale video facial attributes dataset,” in European Conf. on Comp. Vision (ECCV) , pp. 650–667, 2022

    2022

  42. [50]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll´ar, and L. Zitnick, “Microsoft coco: Common objects in context,” in European Conf. on Comp. Vision (ECCV) , pp. 740–755, 2014

    2014

  43. [51]

    Semantics- guided diffusion for deep joint source-channel coding in wireless image transmission,

    M. Zhang, H. Wu, G. Zhu, R. Jin, X. Chen, and D. G ¨und¨uz, “Semantics- guided diffusion for deep joint source-channel coding in wireless image transmission,” arXiv preprint arXiv:2501.01138 , 2025

    2025 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.