arXiv 2409.078962025-03-15cs.CV

MambaMIC：基于状态空间模型的高效显微图像分类基线

MambaMIC: An Efficient Baseline for Microscopic Image Classification with State Space Models

Nanjing Agricultural University（南京农业大学）；National University of Defense Technology（国防科技大学）；Xiangtan University（湘潭大学）；Nanjing University of Posts and Telecommunications（南京邮电大学）；Soochow University（苏州大学）

Shun Zou, Zhuo Zhang, Yi Zou, Guangwei Gao

备注：: 7 pages, 4 figures

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对显微图像分类中全局建模与效率难以兼顾的问题，提出 MambaMIC 双分支状态空间骨干，通过局部卷积、全局 SSM 及特征调制交互模块实现五个数据集上的 SOTA 性能。

摘要：

近年来，基于 CNN 和 Transformer 的方法在显微图像分类（MIC）中取得了显著进展。然而，现有方法仍面临全局建模与高效计算之间的两难。尽管选择性状态空间模型（SSM）能够以线性复杂度模拟长程依赖，但它在 MIC 中仍遇到局部像素遗忘、通道冗余以及局部感知缺失等挑战。为解决这些问题，我们提出了一个简单而高效的 MIC 任务视觉骨干网络，名为 MambaMIC。具体而言，我们引入了一个局部-全局双分支聚合模块：MambaMIC Block，旨在有效捕获并融合局部连接性与全局依赖。在局部分支中，我们使用局部卷积捕获像素相似性，缓解局部像素遗忘并增强感知。在全局分支中，SSM 提取全局依赖，而 Locally Aware Enhanced Filter 减少通道冗余和局部像素遗忘。此外，我们设计了 Feature Modulation Interaction Aggregation Module，用于深层特征交互和关键特征重新定位。大量基准测试表明，MambaMIC 在五个数据集上均达到最先进性能。代码可在 https://zs1314.github.io/MambaMIC 获取。

英文摘要：

In recent years, CNN and Transformer-based methods have made significant progress in Microscopic Image Classification (MIC). However, existing approaches still face the dilemma between global modeling and efficient computation. While the Selective State Space Model (SSM) can simulate long-range dependencies with linear complexity, it still encounters challenges in MIC, such as local pixel forgetting, channel redundancy, and lack of local perception. To address these issues, we propose a simple yet efficient vision backbone for MIC tasks, named MambaMIC. Specifically, we introduce a Local-Global dual-branch aggregation module: the MambaMIC Block, designed to effectively capture and fuse local connectivity and global dependencies. In the local branch, we use local convolutions to capture pixel similarity, mitigating local pixel forgetting and enhancing perception. In the global branch, SSM extracts global dependencies, while Locally Aware Enhanced Filter reduces channel redundancy and local pixel forgetting. Additionally, we design a Feature Modulation Interaction Aggregation Module for deep feature interaction and key feature re-localization. Extensive benchmarking shows that MambaMIC achieves state-of-the-art performance across five datasets. code is available at https://zs1314.github.io/MambaMIC

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。