Search papers, labs, and topics across Lattice.
This paper introduces NeuroRefiner, a multi-agent system designed to enhance 3D neuron segmentation in fluorescence microscopy by mimicking the iterative workflow of human experts. By employing three collaborative agents for diagnosing errors, generating corrections, and validating results, NeuroRefiner effectively addresses the challenges posed by the sparse and elongated morphology of neurons. The method, validated on multiple datasets, achieves a notable 3.02% improvement in F1 score on the ZBFWB dataset compared to existing state-of-the-art techniques, highlighting its superior accuracy and interpretability.
NeuroRefiner's multi-agent approach achieves a 3.02% boost in segmentation accuracy, setting a new benchmark for neuron morphology challenges in microscopy.
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.
