MISTICA: Minimum Spanning Tree-Based Coarse Image Alignment for Microscopy Image Sequences.

Area of Science:

• Biomedical Imaging
• Computational Biology
• Medical Image Analysis

Background:

• In vivo microscopy image registration is crucial for studying diseases like atherosclerosis.
• Cardiac and respiratory motion, focal plane changes, and poor image quality hinder accurate registration.
• Existing coarse alignment methods struggle with long sequences and often require manual intervention.

Purpose of the Study:

• To develop an automated coarse alignment method for in vivo microscopy image sequences.
• To overcome limitations of existing methods in handling poor quality images and long sequences.
• To improve the accuracy and efficiency of image registration for biological studies.

Main Methods:

• Proposed MISTICA (Minimum Weighted Spanning Trees for Image Coarse Alignment), a novel automated method.
• MISTICA reorders images into shorter sequences and down-weights poor quality images.
• The method automatically selects an anchor image, eliminating user dependency and mitigating error propagation.

Main Results:

• MISTICA demonstrates superior performance compared to existing alignment methods on mouse artery microscopy sequences.
• The method successfully addresses challenges posed by long sequences and image quality variations.
• Automated anchor image selection ensures a fully automated and robust registration process.

Conclusions:

• MISTICA provides a computationally efficient and automated solution for coarse alignment of in vivo microscopy image sequences.
• This novel approach enhances the reliability of image registration in complex biological studies.
• MISTICA offers a significant advancement for researchers studying dynamic biological processes in vivo.