MCF

MCF implements model-controlled flooding to integrate prior object models into watershed-based image reconstruction and segmentation for automated single-cell analysis and other microscopic-image applications.

Key Features:

• Integration of A Priori Information: Incorporates predefined or custom model functions that refine initial seeding, region growing, and stopping rules within the flooding process.
• Customizable Simulation: Simulation parameters can be configured using user-defined model functions or default functions to adapt the flooding simulation to different imaging contexts.
• Extension of Connected Attribute Filters: Modifies connected components of grayscale images using operations such as size transform and extensions of grayscale area opening and attribute thickening/thinning for image reconstruction.

Scientific Applications:

• Single Cell Analysis: Applied to automated single-cell studies to support tasks such as speckle counting.
• Concealed Object Detection: Used to identify hidden or concealed objects within complex image datasets.
• Benchmark Microscopic Image Data Sets: Validated on benchmark microscopic image data sets and reported to achieve improved error rates relative to existing algorithms.

Methodology:

Integrates prior knowledge into watershed-transform segmentation via flooding simulations guided by predefined or custom model functions that refine initial seeding, region growing, and stopping rules; modifies connected components with connected attribute filters including size transform, extensions of grayscale area opening, and attribute thickening/thinning; performs a series of transformations and simulations to redefine images for downstream segmentation.