TeraStitcher

TeraStitcher stitches teravoxel-sized tiled 3D microscopy images into spatially registered multiresolution volumes for large-scale microscopy image analysis.

Key Features:

• Efficient Memory Utilization: Minimizes memory usage and processes only necessary portions of teravoxel-sized datasets, enabling operation on workstations with less than 8 GB of RAM.
• Adaptive Algorithms: Employs ad-hoc algorithms for stitching very large images and leverages approximate tile positions to optimize alignment and performance.
• Multiresolution Image Representation: Saves stitched results in a multiresolution format to support efficient retrieval and processing at different spatial scales.
• Versatile Software Architecture: Separates memory-efficient strategies from algorithms that depend on image characteristics, allowing modular replacement of stitching methods.

Scientific Applications:

• Whole Mouse Brain Reconstruction: Stitching teravoxel-sized whole mouse brain 3D microscopy datasets acquired at micrometer resolution.
• Benchmarking and Evaluation: Performance benchmarking against state-of-the-art stitching tools using publicly available megavoxel-sized datasets.

Methodology:

Performs memory-efficient data management and ad-hoc stitching algorithms that exploit known approximate tile positions, outputs multiresolution volumes, and separates memory strategies from image-characteristic-specific algorithms.