Igneous

Igneous processes petascale three-dimensional electron microscopy images to generate dense-segmentation meshes, neuron skeletons, and multi-resolution image hierarchies for neuroanatomical visualization and analysis.

Key Features:

• Distributed Dense 3D Segmentation Meshing: Generates meshes from dense 3D segmentations of brain tissue to produce detailed surface representations.
• Neuron Skeletonization: Derives neuron skeletons from segmentation data to support structural and connectivity analyses.
• Hierarchical Downsampling: Creates multi-resolution image hierarchies through downsampling to enable analysis at different levels of detail.
• Data Management: Manages large volumes of imaging-derived data, supporting storage and retrieval of meshes, skeletons, and image hierarchies.
• Scalability: Scales horizontally across cloud and cluster computing resources to handle petascale datasets.

Scientific Applications:

• Connectomics Reconstruction: Supports reconstruction of neural connectivity by producing segmentation-derived neuron skeletons suitable for connectivity analysis.
• Morphological Visualization and Analysis: Provides multi-resolution meshes and segmentation-derived surfaces for detailed morphological visualization and quantitative analysis.
• Multi-scale Image Exploration and Processing: Enables working with electron microscopy data at multiple resolutions via hierarchical downsampling and image hierarchies.

Methodology:

Igneous uses a Python-based distributed computing framework to parallelize computation across multiple nodes for mass production of meshes, neuron skeletons, and multi-resolution image hierarchies.