NeuroRefiner: Morphology-Aware Multi-Agent Refinement for 3D Fluorescence Microscopy Neuron Segmentation
2026-08-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors created NeuroRefiner, a system that improves 3D images of neurons by mimicking how experts fix mistakes. It uses three parts that work together to find errors, suggest corrections, and check the improved results. They also developed TopoRefineNet, a tool that helps make the neuron shapes more accurate using special image features. Tests show their method works better than others, especially on difficult datasets. Overall, their approach helps make detailed neuron images clearer and less broken.
3D neuron segmentationfluorescence microscopytopological errorsU-Netcross-modality feature fusioniterative refinementBigNeuron datasetF1 scorevoxel-level editingmulti-agent system
Authors
Haiyang Yan, Jinyue Guo, Yanchao Zhang, Bingqing Wang, Zhenchen Li, Jing Liu, Jiazheng Liu, Linlin Li, Hua Han
Abstract
Accurate 3D neuron segmentation in fluorescence microscopy is critical for neuroscience. However, the sparse and elongated morphology of neurons poses significant challenges to existing segmentation methods. These methods struggle to preserve both local details and global topology, leading to fragmented results. To address this, we propose NeuroRefiner, a multi-agent system that formalizes the human expert workflow involving iterative global observation and local editing. Specifically, NeuroRefiner comprises three collaborative agents dedicated to diagnosing topological errors, generating correction instructions, and validating refinement quality. To facilitate agent instruction-guided segmentation refinement, we propose TopoRefineNet, a dedicated 3D U-Net-based tool that leverages cross-modality feature fusion to generate refined masks. Through multi-round agent reasoning and voxel-level editing, NeuroRefiner produces topologically more accurate segmentations with enhanced interpretability. Experiments on the BigNeuron, CWMBS, and ZBFWB datasets demonstrate that NeuroRefiner outperforms state-of-the-art methods, notably achieving a 3.02% improvement in F1 score on the challenging ZBFWB dataset.