文献类型：期刊文献

英文题名：An Improved Low-Bit-Rate Image Compression Framework Based on Semantic-Aware Model and Neighborhood Attention

作者：Zeng, Chengbin Zhang, Liang

第一作者：Zeng, Chengbin

通信作者：Zeng, CB[1]

机构：[1]Moutai Inst, Dept Automat, Renhuai 564507, Guizhou, Peoples R China;[2]Guizhou Inst Technol, Sch Big Data, Guiyang 550025, Guizhou, Peoples R China

第一机构：Moutai Inst, Dept Automat, Renhuai 564507, Guizhou, Peoples R China

通信机构：corresponding author), Moutai Inst, Dept Automat, Renhuai 564507, Guizhou, Peoples R China.

年份：2025

卷号：13

起止页码：139490-139505

外文期刊名：IEEE ACCESS

收录：;EI(收录号：20253318984543);Scopus(收录号：2-s2.0-105013121304);WOS:【SCI-EXPANDED(收录号:WOS:001550816100011)】；

基金：This work was supported in part by the National Natural Science Foundation of China under Grant 61966006, in part by Zunyi Science and Technology Plan Project under Grant Zunshi Kehe HZ Zi (2024) 385, and in part by the Science Research Foundation for High-Level Talents of Moutai Institute under Grant mygccrc[2024]012.

语种：英文

外文关键词：Image compression; semantic-aware model; semantic-aware model; neighborhood attention; neighborhood attention; residual vector quantization; residual vector quantization; residual vector quantization

摘要：In recent years, image compression techniques based on deep neural networks have achieved significant advancements, outperforming traditional methods in delivering higher compression efficiency at lower bit rates. However, existing approaches often result in degraded image quality at low bit rates, leading to distortions in critical regions such as faces and text during decoding. To address these limitations, we propose a robust image compression framework based on a semantic-aware model and a hyper-prior encoder with neighborhood attention. First, we utilize the encoder of the semantic-aware model to transform the input image into a latent space Z. To further improve the information representation within the latent space Z, we design a hyper-prior encoder, which leverages neighborhood attention to perform feature enhancement and transformation. This process can minimize quantization errors and facilitate efficient vector quantization. The refined latent space is then quantized using a residual vector quantization technique to ensure efficient and compact representation. Finally, entropy coding is applied to the quantized latent space, enabling high compression efficiency. Experimental results on public benchmark datasets show that our proposed framework achieves comparable results to current mainstream image compression methods. In addition, our proposed algorithm effectively mitigates distortions in facial and textual regions while preserving the structural integrity and visual fidelity. In future work, we plan to improve the model's robustness under low-light conditions and enhance compression efficiency through lightweight optimization techniques, enabling broader real-world deployment.