MambaLiteUNet：面向鲁棒皮肤病变分割的跨门控自适应特征融合
MambaLiteUNet: Cross-Gated Adaptive Feature Fusion for Robust Skin Lesion Segmentation
把 Mamba 状态空间建模嵌入 U-Net 的跳跃连接，用跨门控注意力与局部-全局特征混合模块在极低参数量下保留细粒度边界，可作为轻量分割网络在皮肤镜等专业成像域的可复用骨干设计。
原文依据 integrates Mamba state space modeling into a U-Net architecture, along with three key modules
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期的分割模型通过大幅减少参数量和计算复杂度，展现出了可观的效率。然而，这些模型往往难以准确勾勒出对皮肤癌早期诊断与治疗方案制定至关重要的细微病变边界与纹理模式。本文提出 MambaLiteUNet，一个紧凑而鲁棒的分割框架，它将 Mamba 状态空间建模集成到 U-Net 架构中，并包含三个关键模块：自适应多分支 Mamba 特征融合（AMF）、局部全局特征混合（LGFM）以及跨门控注意力（CGA）。这些模块旨在增强局部-全局特征交互、保留空间细节并提升跳跃连接的质量。MambaLiteUNet 在 ISIC2017、ISIC2018、HAM10000 和 PH2 四个基准上取得 87.12% 的平均 IoU 和 93.09% 的平均 Dice 分数，优于最优模型。与 U-Net 相比，我们的模型将平均 IoU 和 Dice 分别提升 7.72 和 4.61 个点，同时参数量减少 93.6%、GFLOPs 减少 97.6%。此外，在包含六个未见病变类别的域泛化设定下，MambaLiteUNet 取得 77.61% 的 IoU 和 87.23% 的 Dice，在所有被评估模型中表现最佳。大量实验表明，MambaLiteUNet 在精度与效率之间取得了良好平衡，使其成为皮肤科图像分割中具有竞争力且实用的方案。我们的代码已公开于：https://github.com/maklachur/MambaLiteUNet。
英文原文摘要 1536 字符
Recent segmentation models have demonstrated promising efficiency by aggressively reducing parameter counts and computational complexity. However, these models often struggle to accurately delineate fine lesion boundaries and texture patterns essential for early skin cancer diagnosis and treatment planning. In this paper, we propose MambaLiteUNet, a compact yet robust segmentation framework that integrates Mamba state space modeling into a U-Net architecture, along with three key modules: Adaptive Multi-Branch Mamba Feature Fusion (AMF), Local Global Feature Mixing (LGFM), and Cross-Gated Attention (CGA). These modules are designed to enhance local-global feature interaction, preserve spatial details, and improve the quality of skip connections. MambaLiteUNet achieves an average IoU of 87.12% and average Dice score of 93.09% across ISIC2017, ISIC2018, HAM10000, and PH2 benchmarks, outperforming state-of-the-art models. Compared to U-Net, our model improves average IoU and Dice by 7.72 and 4.61 points, respectively, while reducing parameters by 93.6% and GFLOPs by 97.6%. Additionally, in domain generalization with six unseen lesion categories, MambaLiteUNet achieves 77.61% IoU and 87.23% Dice, performing best among all evaluated models. Our extensive experiments demonstrate that MambaLiteUNet achieves a strong balance between accuracy and efficiency, making it a competitive and practical solution for dermatological image segmentation. Our code is publicly available at: https://github.com/maklachur/MambaLiteUNet.