CellMamba: Adaptive Mamba for Accurate and Efficient Cell Detection

2025-12-25

AI Summary

Pathological image cell detection faces challenges including high cell density, subtle inter-class distinctions, and severe background interference. To address these, we propose a lightweight, efficient one-stage fine-grained detector. Our method introduces the novel Triple-Mapping Adaptive Coupling (TMAC) module, which adaptively fuses local sensitivity with global consistency; an Adaptive Mamba Head for dynamic multi-scale feature weighting; and a VSSD backbone integrating NC-Mamba/MSA with TMAC, enhanced by a learnable multi-scale feature fusion mechanism. Evaluated on CoNSeP and CytoDArk0, our approach surpasses CNN-, Transformer-, and Mamba-based baselines in detection accuracy while reducing model parameters by 32% and inference latency by 41%.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multi-instance/Multi-view Learning
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Cell detection in pathological images presents unique challenges due to densely packed objects, subtle inter-class differences, and severe background clutter. In this paper, we propose CellMamba, a lightweight and accurate one-stage detector tailored for fine-grained biomedical instance detection. Built upon a VSSD backbone, CellMamba integrates CellMamba Blocks, which couple either NC-Mamba or Multi-Head Self-Attention (MSA) with a novel Triple-Mapping Adaptive Coupling (TMAC) module. TMAC enhances spatial discriminability by splitting channels into two parallel branches, equipped with dual idiosyncratic and one consensus attention map, adaptively fused to preserve local sensitivity and global consistency. Furthermore, we design an Adaptive Mamba Head that fuses multi-scale features via learnable weights for robust detection under varying object sizes. Extensive experiments on two public datasets-CoNSeP and CytoDArk0-demonstrate that CellMamba outperforms both CNN-based, Transformer-based, and Mamba-based baselines in accuracy, while significantly reducing model size and inference latency. Our results validate CellMamba as an efficient and effective solution for high-resolution cell detection.

Problem

• Detects densely packed cells in pathological images accurately
• Addresses subtle inter-class differences and severe background clutter
• Achieves efficient detection with reduced model size and latency

Innovation

• CellMamba Blocks with Triple-Mapping Adaptive Coupling
• Adaptive Mamba Head fuses multi-scale features
• Lightweight one-stage detector for biomedical images