ViNe-Seg

ViNe-Seg performs deep-learning-based semi-automatic segmentation of visible neurons in optical functional imaging datasets, enabling detection of neural somata regardless of their activity status and facilitating integration with subsequent analysis steps.

Key Features:

• Activity-Independent Neuron Detection: Detects visible neurons irrespective of their activity status, including in low signal-to-noise ratio datasets.
• Real-Time Segmentation Capability: Supports segmentation during an ongoing experiment.
• Expert Supervision Framework: Enables precise identification of Regions of Interest through semi-automatic segmentation with expert oversight.
• Customizable Model Repository: Provides an array of segmentation models with options for training custom models and sharing them with the community.
• Integrated Analysis Pipeline: Seamlessly integrates subsequent analysis steps following segmentation.

Scientific Applications:

• Neuronal Microcircuit Imaging Analysis: Segments neural somata in optical functional imaging of neuronal microcircuits, addressing the most time-consuming step in such analyses.
• Low Signal-to-Noise Ratio Dataset Processing: Produces accurate segmentation results in challenging datasets where current auto-segmentation tools encounter difficulties.
• Cross-Study Dataset Standardization: Reduces variability introduced by manual segmentation methods to improve dataset consistency across studies.

Methodology:

ViNe-Seg employs deep learning approaches for semi-automatic segmentation of neural somata.