Related Experiment Video
Updated: May 15, 2025
17:06
Co-analysis of Brain Structure and Function using fMRI and Diffusion-weighted Imaging
Published on: November 8, 2012
26.1K
Fibre tracing in biomedical images: An objective comparison between seven algorithms
Youssef Arafat1, Cristina Cuesta-Apausa2, Esther Castellano2
1Department of Computer Science, School of Science and Technology, City St George's, University of London, London, United Kingdom.
Plos One
|April 10, 2025
Summary
A new algorithm, Trace Ridges, accurately traces biomedical fibre-like structures like vasculature and nerves. It outperforms existing methods in accuracy and is nearly as fast as the quickest, offering a valuable tool for biomedical image analysis.
Area of Science:
- Biomedical image analysis
- Computer vision
- Medical imaging
Background:
- Analyzing elongated structures in biomedical images is crucial for understanding processes like angiogenesis and treatment effects.
- Existing methods for tracing vasculature, nerves, and extracellular matrix fibers have limitations.
Purpose of the Study:
- To objectively compare six existing methodologies and one novel approach, Trace Ridges, for tracing fiber-like structures in biomedical images.
- To introduce Trace Ridges as a fast, automatic algorithm for accurate delineation of these structures.
Main Methods:
- Comparison of Edge Detection, CT Fire, Scale Space, Twombli, U-Net, Graph Based, and the novel Trace Ridges algorithm.
- Utilized four distinct biomedical datasets with manual ground truth delineation.
- Evaluated algorithms using three distance error metrics (total, average, maximum) and processing time, with options for pre-processing filters (none, Gaussian, DnCnn).
Main Results:
- No single algorithm excelled across all metrics.
- Trace Ridges achieved the lowest total distance error, indicating superior accuracy.
- Trace Ridges demonstrated competitive processing speed, ranking second only to Edge Detection.
Conclusions:
- Trace Ridges offers a significant advancement in accurately and efficiently tracing biomedical fiber-like structures.
- The study provides a comprehensive comparison to guide the selection of appropriate tracing methodologies.
- The developed algorithm and code are publicly available for research use.
