低带宽网络中面向结构保真的语义传输方法

胡天崴; 张祥瑞; 陈剑; 段昊东; 贾杰

doi:10.11999/JEIT260525

低带宽网络中面向结构保真的语义传输方法

doi: 10.11999/JEIT260525 cstr: 32379.14.JEIT260525

东北大学计算机科学与工程学院沈阳 110819

基金项目: 国家自然科学基金资助项目(No.62132004, No.62572109)

详细信息

作者简介:
胡天崴：等：面向低带宽网络的结构保真语义传输方法

通讯作者:
张祥瑞，　20231984@stu.neu.edu.cn

中图分类号: TN911; TN911.73
计量
- 文章访问数: 12
- HTML全文浏览量: 1
- PDF下载量: 0
- 被引次数: 0
出版历程
- 修回日期: 2026-08-17
- 录用日期: 2026-08-17
- 网络出版日期: 2026-08-25

A Structure-Preserving Semantic Transmission Method for Low-Bandwidth Networks

School of Computer Science and Engineering, Northeastern University, Shenyang 110819

Funds: The National Natural Science Foundation of China (No.62132004, No.62572109)

摘要

摘要: 针对极低带宽条件下传统信源编码在低码率场景中易出现“数字悬崖”效应，且现有语义通信方法难以同时兼顾结构保真、色彩一致性与低时延的问题，该文提出一种面向低带宽网络的结构保真语义传输方法。该方法构建语义—结构—外观多模态解耦表示框架，通过引入对象感知的语义引导编码机制，实现关键区域驱动的自适应码率分配；同时设计确定性双流生成重建模型，以结构约束与语义调制协同引导图像重建，从而降低推理复杂度与时延。实验在Set14、DIV2K数据集及基于USRP的LoRa实际物理链路环境下开展，结果表明，在约0.3 bpp的极低码率条件下，该方法可实现低于60 ms的端到端推理时延，并在主体结构一致性与视觉自然度方面优于传统编码方法。研究结果表明，所提方法在受限带宽条件下能够有效提升传输效率与重建质量，适用于高可靠低时延的边缘视觉通信场景。本文相关代码已公开发布于 ScienceDB和github平台，可通过公开链接访问，永久标识符(DOI)为： 10.57760/sciencedb.33990 。
- 语义通信 /
- 生成式重建 /
- 结构保真 /
- 多模态解耦 /
- LoRa
Abstract: Objective Low-bandwidth visual transmission is essential for edge-intelligence applications such as disaster inspection, underwater exploration, and remote assistance. These scenarios require visual communication systems to simultaneously achieve low bitrate, high structural fidelity, stable color reconstruction, and low end-to-end latency. However, conventional image coding methods suffer from severe quality degradation at extremely low bitrates due to the digital-cliff effect and compression artifacts. Although semantic communication provides a promising solution by transmitting task-relevant representations rather than pixel-level information, existing approaches still face challenges in balancing bitrate efficiency, reconstruction fidelity, and decoding complexity. Therefore, a structure-preserving semantic transmission method, termed Edge-Link, is proposed for low-bandwidth networks. Methods Edge-Link adopts a multimodal decoupled representation framework that separates image information into three complementary streams: a semantic stream, a structural stream, and a low-frequency appearance stream. The semantic stream is extracted using a frozen CLIP encoder to provide global semantic guidance, while the structural stream is obtained from Canny edge information to explicitly preserve object boundaries. A low-resolution color map is introduced as the appearance stream to maintain global color distribution and illumination characteristics with minimal transmission overhead. Furthermore, a FastSAM-based semantic gateway is developed to distinguish foreground objects from background regions, and an object-aware bitrate allocation strategy is designed to prioritize important semantic regions under bandwidth constraints. At the receiver, a deterministic dual-stream generation network based on SPADE is proposed, where structural information provides spatial constraints and semantic features guide the reconstruction process, avoiding the high latency caused by iterative diffusion sampling. A real LoRa communication prototype based on GNU Radio and USRP is also implemented to validate transmission feasibility and robustness under practical wireless conditions. Results and Discussions Experiments are conducted on Set14 and DIV2K to evaluate transmission efficiency, reconstruction quality, perceptual naturalness, and latency. The proposed method achieves an average payload of 10.96 KB for 512×512 images, corresponding to about 0.31 bpp, while keeping the end-to-end inference latency below 60 ms; under a LoRa narrowband link of about 8 kbps, the airtime of a single frame is about 11 s, verifying its feasibility in bandwidth-limited environments (Table 1). The ablation study shows that the object-aware coding strategy improves the bitrate-quality trade-off by reducing the average payload on DIV2K from 126.09 KB to 93.55 KB while preserving better foreground reconstruction quality (Table 2). Compared with JPEG, the proposed method also shows better structural preservation and perceptual quality at low bitrates; for example, on image 0843 with a payload of 15.54 KB, the global SSIM is improved from 0.783 to 0.963, and the subject-region LPIPS is reduced from 0.561 to 0.307 (Table 4). Under the same bitrate constraint, the proposed method further outperforms VQGAN and ControlNet+Canny, achieving PSNR, SSIM, LPIPS, and NIQE values of 23.738, 0.622, 0.223, and 4.570, respectively, indicating a better balance between fidelity and perceptual quality in low-bandwidth semantic reconstruction (Table 5). Conclusions A structure-preserving semantic transmission framework for low-bandwidth networks is presented. By combining multimodal decoupled representation, object-aware bitrate allocation, and deterministic dual-stream reconstruction, the framework balances transmission efficiency, structural fidelity, perceptual quality, and decoding latency. The reported results show that the proposed method is well suited to high-reliability edge visual communication scenarios in which accurate contours, stable color appearance, and efficient inference are simultaneously required. The real-link validation on a LoRa prototype further suggests its practical potential for bandwidth-constrained edge networks.
- semantic communication /
- generative reconstruction /
- structure preservation /
- multimodal decoupling /
- LoRa

HTML全文

图 1 系统架构

下载: 全尺寸图片幻灯片

图 2 GNU Radio中lora协议与USRP传输结合流图

下载: 全尺寸图片幻灯片

图 3 三种方法的各指标对比图

下载: 全尺寸图片幻灯片

图 4 Face.png和Foreman.png的视觉对比结果

下载: 全尺寸图片幻灯片

图 5 Bridge.png在8.60 KB条件下的视觉对比结果

下载: 全尺寸图片幻灯片

图 6 基于FastSAM的主体区域与背景区域划分效果

下载: 全尺寸图片幻灯片

图 7 视觉重建对比结果图

下载: 全尺寸图片幻灯片

图 8 三种方法的指标柱状图

下载: 全尺寸图片幻灯片

图 9 NIQE指标对比图

下载: 全尺寸图片幻灯片

图 10 视觉重建对比结果图

下载: 全尺寸图片幻灯片

表 1 系统端到端性能评估结果

指标	实测值	备注
平均载荷	10.96 KB	Set14平均结果
算法推理时延	<60 ms	RTX 3090 (编码&解码)
空口传输时延	≈11 s	LoRa SF=7(约8 kbps)
平均带宽效率	0.31 bpp	10.96 kb/512²

下载: 导出CSV

表 2 DIV2K 数据集对比结果

策略	平均载荷(KB)	全局SSIM	LPIPS	备注
Global	126.09	0.6350	0.3328	无
Object-Aware	93.55	0.6385	全图：0.380主体：0.223	带宽节约26%

下载: 导出CSV

表 3 Edge-Link 与传统编码方法在典型图像上的对比结果

图像	载荷(KB)	方法	SSIM	LPIPS	NIQE
Face	3.63	Ours	0.633	0.393	8.77
Face	3.63	JPEG	0.442	0.659	21.67
Face	3.63	WebP	0.709	0.373	7.81
Foreman	7.64	Ours	0.828	0.223	5.72
Foreman	7.64	JPEG	0.802	0.334	12.97
Foreman	7.64	WebP	0.925	0.083	6.51
Bridge	8.60	Ours	0.388	0.505	4.26
Bridge	8.60	JPEG	0.596	0.348	7.47
Bridge	8.60	WebP	0.650	0.365	4.90
Average	10.96	Ours	0.716	0.236	5.52
Average	10.96	JPEG	0.696	0.304	8.794
注：NIQE指标越低代表图像越自然。本方法只在背景信息较多的图片中数据表现优于WebP，对于平均数据而言，本方法并不具有优势，故未展示。

下载: 导出CSV

表 4 部分DIV2K 高分辨率图像局部细节重建质量对比

图片	载荷(KB)	方法	整体LPIPS	Subject-ROI LPIPS	整体SSIM	整体NIQE	Subject-ROI NIQE	备注
0843	15.54	Ours	0.059	0.307	0.963	8.13	9.56	极简场景
0843	15.54	JPEG	0.112	0.561	0.783	11.68	10.94	极简场景
0843	15.54	WebP	0.093	0.114	0.970	8.79	7.86	极简场景
0877	36.71	Ours	0.201	0.226	0.874	6.33	8.30	丰富纹理
0877	36.71	JPEG	0.402	0.416	0.791	16.19	21.95	丰富纹理
0877	36.71	WebP	0.065	0.139	0.924	6.97	9.31	丰富纹理
平均	～93	Ours	0.371	0.223	0.622	4.57	7.80	平均数据
平均	～93	JPEG	0.222	0.305	0.739	5.530	10.122	平均数据
注：Dataset Avg.中的 Subject-ROI LPIPS为全集主体区域测量均值，用于评估语义前景区域的重建。对于WebP，本方法在平均数据上并不具有优势，故未展示。

下载: 导出CSV

表 5 不同视觉重建方法的性能比较

方法	PSNR	SSIM	LPIPS	NIQE
Ours	23.738	0.622	0.223	4.570
VQGAN	21.965	0.588	0.220	6.026
ControlNet+Canny	9.618	0.189	0.731	10.071

下载: 导出CSV

表 6 多模态解耦消融实验结果(Set14数据集)

实验配置	整体SSIM	整体LPIPS	主体SSIM	主体LPIPS
完整模型	0.572	0.358	0.716	0.236
去除结构流	0.510	0.522	0.671	0.343
去除语义流	0.537	0.420	0.693	0.278
去除外观流	0.389	0.485	0.464	0.425

下载: 导出CSV

表 7 双流生成架构消融实验对比

方法	主体SSIM	主体LPIPS
本文双流架构	0.716	0.236
全部特征全连接	0.634	0.258

下载: 导出CSV

参考文献(21)

[1]	JERNBERG C, SANDIN J, ZIEMKE T, et al. The effect of latency, speed and task on remote operation of vehicles[J]. Transportation Research Interdisciplinary Perspectives, 2024, 26: 101152. doi: 10.1016/j.trip.2024.101152.
[2]	BOURTSOULATZE E, KURKA D B, and GÜNDÜZ D. Deep joint source-channel coding for wireless image transmission[J]. IEEE Transactions on Cognitive Communications and Networking, 2019, 5(3): 567–579. doi: 10.1109/TCCN.2019.2919300.
[3]	DONG Chao, DENG Yubin, LOY C C, et al. Compression artifacts reduction by a deep convolutional network[C]. Proceedings of the 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, 2015: 576–584. doi: 10.1109/ICCV.2015.73.
[4]	LIU Yating, WANG Xiaojie, NING Zhaolong, et al. A survey on semantic communications: Technologies, solutions, applications and challenges[J]. Digital Communications and Networks, 2024, 10(3): 528–545. doi: 10.1016/j.dcan.2023.05.010.
[5]	CHAI Jingxuan, XIAO Yong, and SHI Guangming. On the rate-distortion-complexity tradeoff for semantic communication[J]. IEEE Internet of Things Journal, 2026, 13(14): 31768–31781. doi: 10.1109/JIOT.2026.3689652.
[6]	陈阳, 马欢, 姬智, 等. 面向图像恢复任务的语义通信网络能耗优化[J]. 电子与信息学报, 2026, 48(1): 183–190. doi: 10.11999/JEIT250915. CHEN Yang, MA Huan, JI Zhi, et al. Optimization of energy consumption in semantic communication networks for image recovery tasks[J]. Journal of Electronics & Information Technology, 2026, 48(1): 183–190. doi: 10.11999/JEIT250915.
[7]	AITHAL S K, MAINI P, LIPTON Z, et al. Understanding hallucinations in diffusion models through mode interpolation[C]. Proceedings of the 38th International Conference on Neural Information Processing Systems, Vancouver, Canada, 2024: 134614–134644. doi: 10.52202/079017-4278.
[8]	JIA Zhaoyang, LI Jiahao, LI Bin, et al. Generative latent coding for ultra-low bitrate image compression[C]. Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, USA, 2024: 26088–26098. doi: 10.1109/CVPR52733.2024.02465.
[9]	PEZONE F, MUSA O, CAIRE G, et al. Semantic-preserving image coding based on conditional diffusion models[C]. ICASSP 2024–2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Seoul, Korea, 2024: 13501–13505. doi: 10.1109/ICASSP48485.2024.10447279.
[10]	ZHANG Chaoning, CHO J, PUSPITASARI F D, et al. A survey on Segment Anything Model (SAM): Vision foundation model meets prompt engineering[EB/OL]. arXiv preprint arXiv: 2306.06211. https://arxiv.org/abs/2306.06211, 2023.
[11]	KINGMA D P and WELLING M. Auto-encoding variational Bayes[EB/OL]. arXiv preprint arXiv: 1312.6114. https://arxiv.org/abs/1312.6114, 2013.
[12]	BURGESS C P, HIGGINS I, PAL A, et al. Understanding disentangling in β-VAE[EB/OL]. arXiv preprint arXiv: 1804.03599. https://arxiv.org/abs/1804.03599?context=cs.LG#1, 2018.
[13]	罗一畅, 齐析屿, 张博锐, 等. 分割一切模型的轻量化研究综述[J]. 电子与信息学报, 2026, 48(2): 713–731. doi: 10.11999/JEIT250894. LUO Yichang, QI Xiyu, ZHANG Borui, et al. A survey of lightweight techniques for segment anything model[J]. Journal of Electronics & Information Technology, 2026, 48(2): 713–731. doi: 10.11999/JEIT250894.
[14]	TAPPAREL J, AFISIADIS O, MAYORAZ P, et al. An open-source LoRa physical layer prototype on GNU radio[C]. Proceedings of the 2020 IEEE 21st International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Atlanta, USA, 2020: 1–5. doi: 10.1109/SPAWC48557.2020.9154273.
[15]	SUN Xiaorui, LIU Jun, SHEN Hengtao, et al. On efficient variants of Segment Anything Model: A survey[J]. International Journal of Computer Vision, 2025, 133(10): 7406–7436. doi: 10.1007/s11263-025-02539-8.
[16]	RADFORD A, KIM J W, HALLACY C, et al. Learning transferable visual models from natural language supervision[C]. Proceedings of the 38th International Conference on Machine Learning, 2021: 8748–8763. (查阅网上资料, 未找见本条文献出版地信息, 请确认).
[17]	PARK T, LIU Mingyu, WANG Tingchun, et al. Semantic image synthesis with spatially-adaptive normalization[C]. Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, USA, 2019: 2332–2341. doi: 10.1109/CVPR.2019.00244.
[18]	BLAU Y and MICHAELI T. The perception-distortion tradeoff[C]. Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, USA, 2018: 6228–6237. doi: 10.1109/CVPR.2018.00652.
[19]	JIANG Wei, ZHAI Yongqi, LI Hangyu, et al. Learned image compression with ROI-weighted distortion and bit allocation[EB/OL]. CoRR. https://arxiv.org/html/2401.08154v2, 2024.
[20]	MAO Qi, YANG Tinghan, ZHANG Yinuo, et al. Extreme image compression using fine-tuned VQGANs[C]. Proceedings of the 2024 Data Compression Conference (DCC), Snowbird, USA, 2024: 203–212. doi: 10.1109/DCC58796.2024.00028.
[21]	CHEN Weilong, XU Wenxuan, CHEN Haoran, et al. Semantic communication based on large language model for underwater image transmission[J]. IEEE Transactions on Mobile Computing, 2026, 25(2): 2060–2075. doi: 10.1109/TMC.2025.3607717.