BIRDS

BIRDS performs adaptive bi-channel image registration and deep-learning segmentation to map and analyze three-dimensional mouse brain microscopy datasets.

Key Features:

• Adaptive Bi-Channel Registration: An adaptive bi-channel registration algorithm improves alignment accuracy across dual-channel whole-brain microscopy datasets.
• Deep-Learning Integration: Neural-network-based deep-learning segmentation is trained on registration-generated data and handles incomplete or defective brain images.
• Comprehensive Pipeline: Pipeline stages include image preprocessing, bi-channel registration, automatic annotation, 3D digital frame creation, high-resolution visualization, and expandable quantitative analysis.
• Versatile Application Modes: Supports registration-based segmentation for cross-modality whole-brain datasets and inference-based real-time segmentation for rapid region-of-interest analysis.

Scientific Applications:

• Comparative Structural Analysis: Comparative analysis of brain structures across different conditions or genetic modifications.
• Neural Circuit Visualization: High-resolution visualization of neural circuits and connectivity patterns.
• Quantitative Morphological Assessment: Quantitative assessment of morphological changes within the mouse brain.

Methodology:

Adaptive bi-channel registration algorithm; image preprocessing; automatic annotation; 3D digital frame construction; neural-network deep-learning segmentation trained on registration-generated training data; registration-based segmentation and inference-based real-time segmentation; high-resolution visualization and expandable quantitative analysis.