Orbit

Orbit performs whole-slide image (WSI) analysis by applying a tile-based map-reduce framework to integrate machine learning and deep learning algorithms for tissue quantification, object detection, and segmentation.

Key Features:

• Tile-Based Processing: Employs a tile-based map-reduce execution framework to divide large WSI files into smaller tiles for scalable processing.
• Machine Learning Integration: Integrates machine learning techniques for tissue quantification in complex biological samples.
• Deep Learning Segmentation: Incorporates deep learning methods for object detection and segmentation within WSIs.
• OMERO Integration: Integrates with the OMERO image server to access and process images from open-source image repositories.
• Distributed Computing Support: Supports scale-out processing via integration with Apache Spark for distributed computing across clusters or cloud infrastructures.

Scientific Applications:

• Idiopathic Lung Fibrosis: Applied to quantify and segment lung tissue in Idiopathic Lung Fibrosis studies.
• Nerve Fiber Density Quantification: Applies deep learning techniques to accurately quantify nerve fiber densities in tissue sections.
• Glomeruli Detection in Kidney: Detects glomeruli in kidney tissue using advanced image analysis algorithms.

Methodology:

Uses a tile-based map-reduce framework to break WSIs into tiles and integrates machine learning and deep learning models, with support for OMERO and Apache Spark for distributed processing.