While deep learning-based image compression has achieved remarkable rate-distortion performance, existing entropy models are often hindered by rigid channel-wise grouping and oversimplified probability distribution assumptions. To address these limitations, this paper proposes a novel entropy modeling framework integrating an Unevenly Grouped Channel Context Model (UGCCM) and a Gaussian Mixture Model (GMM). Specifically, leveraging the information compaction characteristics of latent representations, we design a progressive channel grouping strategy that assigns finer granularity to information-rich initial channels, thereby maximizing contextual information utilization. Furthermore, we replace the conventional single Gaussian prior with a multi-component Gaussian mixture model, significantly enhancing the flexibility and accuracy of latent distribution estimations. Experimental results demonstrate that our framework achieves consistent rate-distortion improvements, yielding over a 0.15 dB PSNR gain at equivalent bitrates on standard test sets compared to baseline models. Moreover, subjective quality assessments confirm our method’s superior capability in preserving complex textures and fine visual details.
Zhiyuan Zhu, Lin Huang, Shi-Hang Ding et al.· 2026 International Conferenc...· 0 citations
Prototype-based knowledge sharing effectively mitigates data and model heterogeneity in federated learning (FL) by exchanging class-level semantic information. However, existing methods typically assume all local prototypes are equally reliable. Consequently, low-quality prototypes from heterogeneous models or dynamic clients can contaminate the global aggregation, leading to a vicious cycle of noise accumulation and performance degradation. To address this, we propose FedLEAF, a Federated Learning framework with server-side proactive Evaluation and clientside Adaptive Fusion. Specifically, the server employs an Adaptive Learning Prototypes (ALP) network to dynamically evaluate prototype reliability and generate learnable aggregation weights, ensuring that highquality prototypes exert a primary influence on the global model. Meanwhile, the client utilizes a Historical Consistency Fusion (HCF) strategy to selectively absorb global knowledge by assessing its consistency with locally maintained historical prototypes. Extensive experiments on standard datasets demonstrate that FedLEAF achieves effective improvements in model accuracy and robustness compared to existing methods.
Zhiyuan Zhu, Si-Yi Deng, Dapeng Wu et al.· 2026 International Conferenc...· 0 citations
