BVLab-analyzer: a unified visual toolbox for terabyte-scale whole-brain cell detection, neuron reconstruction, and vascular network extraction

Abstract

BVLab-Analyzer is a standalone graphical software tool for end-to-end analysis of terabyte-scale whole-brain microscopy images. Operating through a zero-code interface, it handles data import (TIFF, BioImageVision, OME-Zarr), deep learning-based semantic segmentation, 3D skeleton reconstruction, and morphological statistical analysis within a unified pipeline. The software supports three target structures—cells, neurons and vasculature—and exports their morphological reconstructions in standard SWC format. To accelerate training data preparation, Smart Annotation tools provide sub-volume interference isolation, coarse-vessel ray tracing, and skeleton auto-interpolation. For inference, optimized block-wise strategies enable a standard 3D U-Net to predict whole-brain volumes up to 12 TB on a single consumer-grade GPU, while an interrupt-and-resume engine ensures fault tolerance during extended computations. By integrating these features, BVLab-Analyzer lowers the hardware barrier for whole-brain analysis and unifies previously fragmented toolchains—from annotation and segmentation to reconstruction and quantification—into a single, accessible software environment.