Pith. sign in

REVIEW 4 major objections 4 minor 51 references

Visual Fidelity Index for Generative Semantic Communications with Critical Information Embedding

T0 review · 4 major / 4 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 The paper defines GVIF, a per-image fidelity measure based on mutual information between received and reference features, and shows that maximizing it adapts semantic filtering and source coding to the channel, beating benchmark schemes.

desk verdict Strong systems story, but GVIF's key scaling assumption (Eq. 21) is unproven and likely false as stated; send it to review and make the authors defend or redraw the metric. read the letter →

arxiv 2505.10405 v1 pith:WYFH7ATX submitted 2025-05-15 eess.IV cs.AIcs.CVcs.LG

Jianhao Huang , Qunsong Zeng , Kaibin Huang This is my paper · ORCID
classification eess.IVcs.AIcs.CVcs.LG
keywords generativesemanticcommunicationvisualfidelitymetricmutualinformationfilteringclassactivationmappingvariationalautoencoderdiffusionmodelchannel-adaptivetransmission
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 proposes a hybrid generative semantic communication system that sends a short text prompt plus only the image features most relevant to the semantic label, then reconstructs the image with a diffusion model at the receiver. To decide which features are worth sending and how much to compress them, it introduces the generative visual information fidelity (GVIF) index, a per-image number between 0 and 1 that measures how much of the original image's feature information survives transmission. GVIF is the ratio of the mutual information between the received features and the reference features to the mutual information of the reference features with itself, both measured under a Gaussian-channel model of human visual processing. The paper shows that maximizing GVIF over the filtering threshold and the source coder yields a channel-adaptive system with higher PSNR on critical regions and lower FID than benchmark schemes. The reason to care is that existing metrics either assess whole distributions, like FID, or ignore semantic content, like PSNR, while GVIF ties visual fidelity directly to the amount of semantically critical information actually transmitted.

What carries the argument

The load-bearing object is the GVIF identity of Eq. (27), which combines four statistical models into a single ratio. Reference image features follow a Gaussian scale mixture $y^r_{ijc}=\theta^r_{ijc}u_{ijc}$ with $u_{ijc}\sim\mathcal{N}(0,1)$, coming from the VAE hyperprior; any lossy source coder is assumed to scale these features by $\beta_{ijc}$; semantic filtering sets unselected features to zero; and the diffusion decoder preserves selected features while generating unselected ones independently. The human visual system is modeled as an additive Gaussian channel with noise variance $\gamma^2$, so the conditional mutual information between distorted and reference features has a closed form. GVIF is then used as the optimization objective: a zero-order gradient method tunes the filtering threshold $\alpha$ for each fixed coder, and the best pre-trained coder is selected under a latency constraint.

What would settle it

Compute, over many images, the residual $y^c_{ijc}-(\bar\theta^c_{ijc}/\bar\theta^r_{ijc})y^r_{ijc}$ between the features produced by a lossy VAE coder and those from the reference coder; if the residual variance is comparable to or larger than the visual-noise variance $\gamma^2$, the scalar-scaling assumption behind Eq. (21) fails and the GVIF numerator in Eq. (43) overstates the shared information. A second check is to regenerate the same transmitted mask and prompt with different diffusion noise samples; if GVIF varies substantially across runs, the single-sample ergodicity assumption used to compute the mutual information is not valid.

Watch

Extended reading notes

Core claim

The central claim is that the visual fidelity of a single generated image can be quantified by $$V(\bar{\boldsymbol{\$\beta$}},\mathcal{P};x)=\frac{\sum_{(i,j,c)\in\mathcal{P}}\log_2\!\left(1+\frac{(\bar{\$\beta$}_{ijc}\bar{\$\theta$}^r_{ijc})^2}{\$gamma^{2}$}\right)}{\sum_{(i,j,c)\in\mathcal{U}}\log_2\!\left(1+\frac{(\bar{\$\theta$}^r_{ijc})^2}{\$gamma^{2}$}\right)},$$ where $\bar{\theta}^r_{ijc}$ are the feature scales extracted by a high-quality reference VAE coder, $\bar{\beta}_{ijc}$ are per-feature scaling factors from the lossy coder, $\mathcal{P}$ is the set of transmitted feature positions, $\mathcal{U}$ is the full feature grid, and $\gamma^2$ is the variance of the additive Gaussian noise used to model the human visual system. The ratio lies in $[0,1]$, equals one for an unaltered image, and grows when the selected feature set contains high-variance, semantically important positions. The paper further claims that using this ratio as the objective for choosing the filtering threshold $\alpha$ and the source coder $\Phi$ produces a channel-adaptive hybrid system that beats a classic image codec and a VAE-only scheme in mask PSNR and FID, and that GVIF correlates with both mask PSNR and the volume of transmitted critical information.

Load-bearing premise

The index assumes that lower-quality encoding only scales down each feature of the reference image by a fixed factor, leaving the underlying content unchanged; if encoding changes the content that a feature represents, the whole fidelity calculation is measuring the wrong thing.

Editorial extensions

If this is right

  • GVIF provides a per-sample fidelity score that increases with both the number of transmitted critical features and the quality of their compression, so it can serve as the objective in other hybrid semantic communication designs.
  • At low receive SNR the optimized threshold rises to protect only the semantically important features, avoiding the sharp performance cliff that full-feature VAE transmission exhibits below roughly 10 dB.
  • Because GVIF is computed from encoder-side feature statistics rather than from the final pixels, it can be evaluated before the diffusion decoder runs, allowing online adaptation of transmission parameters.
  • The experimental trend that FID falls sharply once a small number of critical features are transmitted supports the practical claim that most visual information is concentrated in semantically important regions.

Reading between the lines

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

  • A natural extension is to turn GVIF into a differentiable training loss for the source encoder and the filtering mask, since the index is built from encoder-controlled feature statistics; the paper itself only uses it to select among pre-trained coders and a scalar threshold.
  • The scalar-scaling distortion model suggests a testable generalization: replacing $\beta_{ijc}$ with a learned linear transform between coder feature spaces would keep the index meaningful for encoders whose latent spaces do not align, without retraining a reference coder.
  • The single-sample ergodicity assumption implies that GVIF should be stable when the same transmitted features are decoded with different diffusion noise samples; measuring that stability would let GVIF be reported with confidence intervals.
  • Reading the numerator as preserved 'visual bits' suggests a rate-perceptual-information formulation, connecting GVIF to rate-distortion theory in a way the paper does not spell out.
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

4 major / 4 minor

Summary. The paper develops a hybrid generative semantic communication system in which a text prompt and a subset of semantically critical VAE features are transmitted, with the receiver using a diffusion-based inpainting model for reconstruction. The main technical contributions are a CAM-based semantic filtering mechanism controlled by a threshold alpha, and a proposed metric, GVIF, defined in Eq. (27) as the ratio of conditional mutual information between distorted and reference image features under a Gaussian scale mixture model and an HVS Gaussian channel model. The authors also formulate a channel-adaptive optimization problem that jointly selects the filtering threshold and the VAE source coder by maximizing the average GVIF under a latency constraint. Experiments on ImageNet report that GVIF increases with mask PSNR and transmitted critical information volume, and that the optimized system achieves higher mask PSNR and lower FID than the JPEG2000 and VAE-based benchmarks.

Significance. If the proposed metric is valid, the paper would provide a principled, information-theoretic per-sample fidelity measure for hybrid Gen-SemCom, filling a real gap since standard metrics such as PSNR and FID do not directly relate visual quality to the volume of semantically critical transmitted information. The system-level contribution, namely channel-adaptive joint optimization of semantic filtering and source coding, is plausible and well motivated, and the experimental comparison against JPEG2000, VAE-only, and prompt-only generation is useful. The paper also includes derivations in Appendix B and some empirical validation in Appendix A, which are strengths. However, the mutual-information interpretation of GVIF rests on an unvalidated scaling assumption between separately trained VAE encoders, and the computation procedure never actually measures features of the generated image; these issues directly affect the central claim and must be resolved before the metric can be accepted as stated.

major comments (4)
  1. [§IV-A2, Eq. (21)] The scaling model y^c_ijc = β_ijc y^r_ijc is introduced without derivation, and it is load-bearing for the GVIF claim: Proposition 4.1 evaluates conditional mutual information only because the same latent realization u_ijc is assumed to appear in both y^c and y^r. This assumption does not follow from the GSM model, since the reference and lossy encoders are separately trained neural networks and no mechanism forces their features to share the same underlying random variable u; scaling of variances alone, which motivates Eq. (30), does not imply elementwise scaling of realizations. The paper validates Eqs. (23)–(24) in Appendix A, but provides no validation of Eq. (21). A direct per-element correlation or cosine-similarity analysis between y^c and y^r across the coder family is needed to support the mutual-information interpretation; without it, the closed form in Eq. (27) is not established as the mutual information it claims to be.
  2. [§IV-B2, Eqs. (28)–(30)] The computation procedure for GVIF never evaluates features of the generated image ~x: θ̄^r and θ̄^c are computed from the reference and lossy source encoders, β̄ is the ratio of their standard deviations, and the metric is assembled from these quantities together with the set P. The actual diffusion output enters only through the assumed independence in Eq. (24). Consequently GVIF is a predicted fidelity based on the source coder and the filter, not a measurement of the image that the receiver actually produced. The Appendix A validation of Eqs. (23)–(24) does not close this gap, because that validation concerns y^g while the computation in Eqs. (28)–(30) bypasses y^g entirely. The paper should either compute GVIF from the generated features or explicitly re-frame GVIF as a model-based prediction and validate that prediction against human or established perceptual scores.
  3. [Appendix A, Eq. (24)] The independence of y^g_ijc outside the transmitted set P is validated only through Pearson correlation coefficients in Fig. 12. Zero correlation does not imply zero mutual information, and Eq. (43) requires I(g^d_ijc; y^r_ijc | θ̄^r, β̄) = 0 for every out-of-P element; the proof collapses if the generated features are merely uncorrelated but dependent. A conditional-Gaussian argument or a direct mutual-information estimate is needed. In addition, the appendix writes y = F(x; Φ_D), which appears to be a typo for the encoder parameter Φ_E, and should be corrected.
  4. [§IV-B and §VI-B] The text presents the monotonicity of GVIF with respect to the transmitted set P and the scaling β̄ as a validated property, but this behavior is built into Eq. (27): enlarging P adds positive log terms, and β̄ enters the numerator multiplicatively. The correlation between GVIF and mask PSNR reported in Fig. 7 is therefore partly expected, since both are driven by the same source-coder parameters and transmitted feature count. The independent evidence in Fig. 11(b), namely FID decreasing with GVIF, is valuable and should be emphasized, but the paper should avoid claiming that the monotonicity experiments alone validate the metric's perceptual sensitivity; a comparison against human ratings or established full-reference metrics would strengthen the claim.
minor comments (4)
  1. [Eq. (19)] The index in the universal quantifier is written as y∈[Hy], which should presumably be j∈[Hy]; this typo appears in the definition of the GSM model.
  2. [§IV-B2] The reference coder is defined as Φr = {Φr_E, Φr_D, Ωc_1, Ωr_2}, where the third element appears to be a typo and should be Ωr_1; please check the notation consistency.
  3. [Eq. (8)] The symbol B is used both for the total number of encoded bits and for the channel bandwidth in the latency expression; this overloading may confuse readers and should be disambiguated.
  4. [Fig. 11] The text says 'FID score is a monotonously decreasing function of the GVIF', but the plot appears to show a decreasing relationship; consider using a more precise term such as 'monotonically decreasing' and clarifying whether the claim is empirical rather than theoretical.

Circularity Check

2 steps flagged · score 4.0 of 10

GVIF's sensitivity validation is partly built into its definition; the system-level conclusions rest on external metrics.

  1. self definitional [Section IV-B (Eq. (27)) and Section VI-B (Fig. 9)]
    "The GVIF lies in the interval [0, 1] and is a monotonically increasing function w.r.t. the volume of set P and{ ¯βijc}. ... Experimental results validate the GVIF metric's sensitivity to visual fidelity, correlating with both the PSNR and critical information volume."

    By Eq. (27), the numerator is a sum over P of log2(1+(β̄ijc θ̄r_ijc)^2/γ^2), and by Eq. (22) features outside P are zeroed. Hence GVIF is monotonically increasing in |P| and β by construction. Section VI-B then reports the monotone dependence of average GVIF on mask PSNR (a proxy for β) and on 1−α (a proxy for |P|) as evidence that 'the GVIF metric comprehensively evaluates the content differences.' This confirms an algebraic property of the definition rather than an independent validation of visual-fidelity sensitivity.

  2. other [Section V-A (Eq. (31)) and Section VI-A/VI-C (Fig. 9)]
    "To achieve this, we optimize the set P and the scaling variables ¯β to maximize the expected GVIF for a given receive SNR ... The proposed GVIF metric is used to quantitatively track the performance of the Gen-SemCom system."

    Average GVIF is both the objective being maximized in problem (31) and the metric used to report the 'performance' of the optimized system in Fig. 9. A system optimized to maximize GVIF will, by construction, score highly on GVIF, so the GVIF-vs-SNR comparison with baselines is not independent evidence for the metric's quality. The paper's separate mask-PSNR and FID comparisons are independent external metrics, so this is a secondary and partial circularity.

full rationale

Most of the derivation is self-contained: Proposition 4.1 is a closed-form evaluation of mutual information under the stated GSM, scaling, generation, and HVS models (Eqs. (19)-(27)), and Appendix A provides an independent empirical check of the generation model (23)-(24) using Pearson correlations. The external benchmarks (mask PSNR and FID) are not derived from GVIF, so the central system-performance claim has independent content. The circularity is limited to (i) presenting the formal monotonicity of Eq. (27) in |P| and β as an experimental validation of fidelity sensitivity, and (ii) using the same GVIF as both the optimization objective and one of the reported performance metrics. The scaling ansatz in Eq. (21) is a strong load-bearing assumption that is not validated for equal latent realizations, but that is an assumption/correctness concern rather than a circular reduction. Overall, the partial circularity warrants a score of 4.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central derivation rests on five main assumptions. The GSM model and Gaussian HVS channel are standard from prior image quality assessment. The most fragile are the ad hoc scaling model in Eq. (21) and the independence of generated features outside P in Eq. (24); the latter is only weakly validated via correlations. No new physical entities are introduced.

free parameters (1)
  • γ² = 0.1
    Variance of visual noise in the HVS model; taken from Sheikh and Bovik [34] rather than fitted to the VAE feature domain. It affects all GVIF values and the resulting optimization decisions.
assumptions (5)
  • standard math Image features follow a Gaussian Scale Mixture model: y_ijc = θ_ijc · u_ijc with u ~ N(0,1).
    Used in Section IV-A1, Eq. (19); inherited from the hyperprior image compression literature [10].
  • ad hoc to paper Features from different VAE coders are related by a scalar scaling: y^c_ijc = β_ijc y^r_ijc with shared latent u.
    Introduced in Section IV-A2, Eq. (21); no derivation or experimental test is provided, and the GVIF formula depends on it.
  • domain assumption Generated features outside the transmitted set P are statistically independent of the original features.
    Assumed in Section IV-A2, Eqs. (23)-(24); Appendix A reports low Pearson correlations, which is weaker evidence than actual independence.
  • domain assumption The HVS can be modeled as adding i.i.d. Gaussian noise of variance γ² to each feature.
    Taken from the classical VIF metric [34], applied to the feature domain in Section IV-A3.
  • domain assumption Spatial consistency of CNN features allows the CAM importance matrix to transfer to VAE features.
    Section III-A relies on this to define the filtering set P that drives both GVIF and the transmission strategy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Visual Fidelity Index for Generative Semantic Communications with Critical Information Embedding." pith.science (2026). https://pith.science/paper/WYFH7ATX

@misc{pith2026250510405,
  author       = {Pith},
  title        = {Pith review of: Visual Fidelity Index for Generative Semantic Communications with Critical Information Embedding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WYFH7ATX}},
  note         = {Machine review of arXiv:2505.10405}
}
read the original abstract

Generative semantic communication (Gen-SemCom) with large artificial intelligence (AI) model promises a transformative paradigm for 6G networks, which reduces communication costs by transmitting low-dimensional prompts rather than raw data. However, purely prompt-driven generation loses fine-grained visual details. Additionally, there is a lack of systematic metrics to evaluate the performance of Gen-SemCom systems. To address these issues, we develop a hybrid Gen-SemCom system with a critical information embedding (CIE) framework, where both text prompts and semantically critical features are extracted for transmissions. First, a novel approach of semantic filtering is proposed to select and transmit the semantically critical features of images relevant to semantic label. By integrating the text prompt and critical features, the receiver reconstructs high-fidelity images using a diffusion-based generative model. Next, we propose the generative visual information fidelity (GVIF) metric to evaluate the visual quality of the generated image. By characterizing the statistical models of image features, the GVIF metric quantifies the mutual information between the distorted features and their original counterparts. By maximizing the GVIF metric, we design a channel-adaptive Gen-SemCom system that adaptively control the volume of features and compression rate according to the channel state. Experimental results validate the GVIF metric's sensitivity to visual fidelity, correlating with both the PSNR and critical information volume. In addition, the optimized system achieves superior performance over benchmarking schemes in terms of higher PSNR and lower FID scores.

Figures

Figures reproduced from arXiv: 2505.10405 by the authors.

Figure 1
Figure 1. Comparisons of multi-level SemCom systems with [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The hybrid Gen-SemCom system with the CIE process. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The system operations of the CIE process. Conv, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Semantic importance modeling using the CAM. In this example, we blend the importance matrix with the resized [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Computation procedures of the GVIF metric. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Illustrations of the GVIF’s ability for image content [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: The performance of the GVIF metric with different compression distortions and filtering thresholds. [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Image examples of the proposed Gen-SemCom system with different threshold [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Performance of the proposed scheme with channel [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Image examples of the performance comparisons with different SNRs. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 12
Figure 12. Figure 12: Illustrations of Pearson correlation coefficients over [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 40 canonical work pages

  1. [1]

    Beyond transmitting bits: Context, seman- tics, 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, seman- tics, and task-oriented communications,” IEEE J. Sel. Areas Commun. , vol. 41, no. 1, pp. 5–41, Jan. 2022

    work page 2022

  2. [2]

    Toward wisdom-evolutionary and primitive- concise 6G: A new paradigm of semantic communication networks,

    P. Zhang, W. Xu, H. Gao, K. Niu, X. Xu, X. Qin, C. Yuan, Z. Qin, H. Zhao, J. Wei et al. , “Toward wisdom-evolutionary and primitive- concise 6G: A new paradigm of semantic communication networks,” Eng., vol. 8, pp. 60–73, Jan. 2022

    work page 2022

  3. [3]

    Semantic knowledge base-enabled zero-shot multi-level feature transmission optimization,

    Y . Sun, H. Chen, X. Xu, P. Zhang, and S. Cui, “Semantic knowledge base-enabled zero-shot multi-level feature transmission optimization,” IEEE Trans. Wirel. Commun., vol. 23, no. 5, pp. 4904–4917, May 2024

    work page 2024

  4. [4]

    A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,

    W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,” IEEE Netw., vol. 34, no. 3, pp. 134–142, June 2019

    work page 2019

  5. [5]

    Toward an intelligent edge: Wireless communication meets machine learning,

    G. Zhu, D. Liu, Y . Du, C. You, J. Zhang, and K. Huang, “Toward an intelligent edge: Wireless communication meets machine learning,” IEEE Commun. Mag. , vol. 58, no. 1, pp. 19–25, Jan. 2020

    2020

  6. [6]

    PartialLoading: User scheduling and bandwidth allocation for parameter-sharing edge inference,

    G. Qu, Q. Chen, X. Chen, K. Huang, and Y . Fang, “PartialLoading: User scheduling and bandwidth allocation for parameter-sharing edge inference,” arXiv preprint arXiv:2503.22982 , 2025

    arXiv 2025

  7. [7]

    Split learning in 6g edge networks,

    Z. Lin, G. Qu, X. Chen, and K. Huang, “Split learning in 6g edge networks,” IEEE Wirel. Commun. , vol. 31, no. 4, pp. 170–176, Aug. 2024

    work page 2024

  8. [8]

    Efficient parallel split learning over resource-constrained wireless edge networks,

    Z. Lin, G. Zhu, Y . Deng, X. Chen, Y . Gao, K. Huang, and Y . Fang, “Efficient parallel split learning over resource-constrained wireless edge networks,” IEEE Trans. Mob. Comput., vol. 23, no. 10, pp. 9224–9239, Oct. 2024

    work page 2024

Show all 51 references
  1. [9]

    End-to-end optimized image compression,

    J. Ball ´e, V . Laparra, and E. P. Simoncelli, “End-to-end optimized image compression,” in Proc. Int. Conf. Learn. Repres. (ICLR), Toulon, France, Apr. 2017

    2017

  2. [10]

    Variational image compression with a scale hyperprior,

    J. Ball ´e, D. Minnen, S. Singh, S. J. Hwang, and N. Johnston, “Variational image compression with a scale hyperprior,” in Proc. Int. Conf. Learn. Repres. (ICLR), Vancouver, CA, May 2018

    2018

  3. [11]

    Learned image compression with mixed transformer-cnn architectures,

    J. Liu, H. Sun, and J. Katto, “Learned image compression with mixed transformer-cnn architectures,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2023, pp. 14 388–14 397

    2023

  4. [12]

    Joint task and data-oriented semantic communications: A deep separate source-channel coding scheme,

    J. Huang, D. Li, C. Huang, X. Qin, and W. Zhang, “Joint task and data-oriented semantic communications: A deep separate source-channel coding scheme,” IEEE Internet Things J., vol. 11, no. 2, pp. 2255–2272, Jan. 2024

    2024

  5. [13]

    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, June 2022

    2022

  6. [14]

    Fun- damental limitation of semantic communications: Neural estimation for rate-distortion,

    D. Li, J. Huang, C. Huang, X. Qin, H. Zhang, and P. Zhang, “Fun- damental limitation of semantic communications: Neural estimation for rate-distortion,” J. Commun. Inf. Net. , vol. 8, no. 4, pp. 303–318, Dec. 2023

    2023

  7. [15]

    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. Cog. Commun. Net., vol. 5, no. 3, pp. 567–579, May 2019

    2019

  8. [16]

    D 2-JSCC: Digital deep joint source-channel coding for semantic communications,

    J. Huang, K. Yuan, C. Huang, and K. Huang, “D 2-JSCC: Digital deep joint source-channel coding for semantic communications,” IEEE J. Sel. Areas Commun., vol. 43, no. 4, pp. 1246–1261, Apr. 2025

    2025

  9. [17]

    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, June 2023

    2023

  10. [18]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of the IEEE conference on computer vision and pattern recognition, June 2018, pp. 586–595

    2018

  11. [19]

    Lamosc: Large language model-driven semantic communication system for visual trans- mission,

    Y . Zhao, Y . Yue, S. Hou, B. Cheng, and Y . Huang, “Lamosc: Large language model-driven semantic communication system for visual trans- mission,” IEEE Trans. Cogn. Commun. Netw. , vol. 10, no. 6, pp. 2005– 2018, Dec. 2024

    2005

  12. [20]

    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

  13. [21]

    Generative semantic communication: Diffusion models beyond bit recovery,

    E. Grassucci, S. Barbarossa, and D. Comminiello, “Generative semantic communication: Diffusion models beyond bit recovery,” arXiv preprint arXiv:2306.04321, 2023

    2023 arXiv

  14. [22]

    Large generative model-assisted talking-face semantic communication system,

    F. Jiang, S. Tu, L. Dong, C. Pan, J. Wang, and X. You, “Large generative model-assisted talking-face semantic communication system,” arXiv preprint arXiv:2411.03876 , 2024

    2024 arXiv

  15. [23]

    M4sc: An mllm-based multi-modal, multi-task and multi-user semantic communi- cation system,

    F. Jiang, S. Tu, L. Dong, K. Wang, K. Yang, and C. Pan, “M4sc: An mllm-based multi-modal, multi-task and multi-user semantic communi- cation system,” arXiv preprint arXiv:2502.16418 , 2025

    2025 arXiv

  16. [24]

    Lambo: Large AI model empowered edge intelligence,

    L. Dong, F. Jiang, Y . Peng, K. Wang, K. Yang, C. Pan, and R. Schober, “Lambo: Large AI model empowered edge intelligence,”IEEE Commun. Mag., vol. 63, no. 4, pp. 88–94, 2025

    2025

  17. [25]

    Hsplitlora: A heterogeneous split parameter- efficient fine-tuning framework for large language models,

    Z. Lin, Y . Zhang, Z. Chen, Z. Fang, X. Chen, P. Vepakomma, W. Ni, J. Luo, and Y . Gao, “Hsplitlora: A heterogeneous split parameter- efficient fine-tuning framework for large language models,” arXiv preprint arXiv:2505.02795, 2025

    2025 arXiv

  18. [26]

    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 Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., June 2022, pp. 10 684– 10 695

    2022

  19. [27]

    A survey of large language models,

    W. X. Zhao, K. Zhou, J. Li, T. Tang, X. Wang, Y . Hou, Y . Min, B. Zhang, J. Zhang, Z. Dong et al. , “A survey of large language models,” arXiv preprint arXiv:2303.18223, 2023

    2023 arXiv

  20. [28]

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

    M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, and S. Hochreiter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” Adv. Neural Inf. Process. Syst. , vol. 30, pp. 6626–6637, Dec. 2017

    2017

  21. [29]

    Generative network layer for communication systems with artificial intelligence,

    M. Thorsager, I. Leyva-Mayorga, B. Soret, and P. Popovski, “Generative network layer for communication systems with artificial intelligence,” IEEE Networking Letters , pp. 82–86, Jan. 2024

    2024

  22. [30]

    Generative ai-aided joint training-free secure semantic communications via multi-modal prompts,

    H. Du, G. Liu, D. Niyato, J. Zhang, J. Kang, Z. Xiong, B. Ai, and D. I. Kim, “Generative ai-aided joint training-free secure semantic communications via multi-modal prompts,” in Proc. IEEE Int. Conf. Acoust., Speech, Signal Process. (ICASSP). IEEE, April 2024, pp. 12 896–12 900

    2024

  23. [31]

    Segment anything meets semantic communication,

    S. Tariq, B. E. Arfeto, C. Zhang, and H. Shin, “Segment anything meets semantic communication,” arXiv preprint arXiv:2306.02094 , 2023

    2023 arXiv

  24. [32]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,” in Proc. of the IEEE/CVF Int. Conf. on Computer Vision (ICCV) , Oct. 2023, pp. 4015–4026

    2023

  25. [33]

    Learning deep features for discriminative localization,

    B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, “Learning deep features for discriminative localization,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , 2016, pp. 2921–2929

    2016

  26. [34]

    Image information and visual quality,

    H. R. Sheikh and A. C. Bovik, “Image information and visual quality,” IEEE Trans. on Image Processing , vol. 15, no. 2, pp. 430–444, 2006

    2006

  27. [35]

    ChatGPT (Mar 14 version) [Large language model],

    OpenAI, “ChatGPT (Mar 14 version) [Large language model],” 2023, accessed: Mar. 14, 2023. [Online]. Available: https://chat.openai.com

    2023

  28. [36]

    Mobile edge intelligence for large language models: A contemporary survey,

    G. Qu, Q. Chen, W. Wei, Z. Lin, X. Chen, and K. Huang, “Mobile edge intelligence for large language models: A contemporary survey,” Early Access in IEEE Commun. Surv. Tutor. , 2025

    2025

  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. Int. Conf. Mach. Learn. (ICML) , Honolulu, USA, 2023, pp. 19 730–19 742

    2023

  30. [38]

    Arithmetic coding for data compression,

    I. H. Witten, R. M. Neal, and J. G. Cleary, “Arithmetic coding for data compression,” Commun. ACM, vol. 30, no. 6, pp. 520–540, 1987

    1987

  31. [39]

    Tse and P

    D. Tse and P. Viswanath, Fundamentals of wireless communication . Cambridge University Press, 2005

    2005

  32. [40]

    Repaint: Inpainting using denoising diffusion probabilistic models,

    A. Lugmayr, M. Danelljan, A. Romero, F. Yu, R. Timofte, and L. Van Gool, “Repaint: Inpainting using denoising diffusion probabilistic models,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 2022, pp. 11 461–11 471

    2022

  33. [41]

    R. G. Gallager, Information theory and reliable communication . New York, NY , USA: Wiley, 1968

    1968

  34. [42]

    Spatially consistent representation learning,

    B. Roh, W. Shin, I. Kim, and S. Kim, “Spatially consistent representation learning,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., June 2021, pp. 1144–1153

    2021

  35. [43]

    On runlength codes,

    E. Zehavi and J. K. Wolf, “On runlength codes,” IEEE Trans. on Inf. Theory, vol. 34, no. 1, pp. 45–54, Aug. 1988

    1988

  36. [44]

    An information fidelity criterion for image quality assessment using natural scene statistics,

    H. R. Sheikh, A. C. Bovik, and G. De Veciana, “An information fidelity criterion for image quality assessment using natural scene statistics,” IEEE Trans. on Image Processing, vol. 14, no. 12, pp. 2117–2128, Nov. 2005

    2005

  37. [45]

    Boyd and L

    S. Boyd and L. Vandenberghe, Convex optimization. UK: Cambridge University Press, 2004

    2004

  38. [46]

    A primer on zeroth-order optimization in signal processing and machine learning: Principals, recent advances, and applications,

    S. Liu, P.-Y . Chen, B. Kailkhura, G. Zhang, A. O. Hero III, and P. K. Varshney, “A primer on zeroth-order optimization in signal processing and machine learning: Principals, recent advances, and applications,” IEEE Signal Process. Mag. , vol. 37, no. 5, pp. 43–54, Sept. 2020

    2020

  39. [47]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in IEEE Conf. Comput. Vis. Pattern Recog. (CVPR) , June 2016, pp. 770–778

    2016

  40. [48]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, L. Kai, and F.-F. Li, “Imagenet: A large-scale hierarchical image database,” in 2009 IEEE Conf. Comput. Vis. Pattern Recog. (CVPR), Miami, FL, USA, June 2009, pp. 248–255

    2009

  41. [49]

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

    A. Kuznetsova, H. Rom, N. Alldrin, J. Uijlings, I. Krasin, J. Pont- Tuset, S. Kamali, S. Popov, M. Malloci, A. Kolesnikov et al. , “The open images dataset v4: Unified image classification, object detection, and visual relationship detection at scale,” Inter. J. of Comput. Vis...

    1956

  42. [50]

    The JPEG2000 still image coding system: an overview,

    C. Christopoulos, A. Skodras, and T. Ebrahimi, “The JPEG2000 still image coding system: an overview,” IEEE Trans. Cons. Elec. , vol. 46, no. 4, pp. 1103–1127, Nov. 2000

    2000

  43. [51]

    BPG image format

    F. Bellard, “BPG image format”,” 2014. [Online]. Available: http://bellard.org/bpg/

    2014

Pith tools

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