Related Experiment Video
Updated: Jul 18, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
Traditional and deep learning-oriented medical and biological image analysis
Bratislavske Lekarske Listy
|August 28, 2023
Summary
This study enhances medical and biological image segmentation using advanced techniques like GrabCut, fuzzy logic, and deep learning (U-Net). These methods improve diagnostic accuracy and cell analysis by providing sharper image boundaries.
Area of Science:
- Medical Imaging
- Bioinformatics
- Computational Biology
Background:
- Accurate segmentation of medical and biological images is crucial for diagnostics and research.
- Existing methods face challenges in segmenting complex cellular structures and medical data.
Purpose of the Study:
- To investigate and improve image segmentation techniques for medical and biological data.
- To enhance diagnostic processes and cell/iron diagnostics through advanced image analysis.
- To introduce novel software and mathematical approaches for superior segmentation results.
Main Methods:
- Implementation of the GrabCut algorithm using C++.
- Development of a fuzzy approach and fuzzy processing for tissue analysis in Matlab.
- Application of deep learning with a U-Net architecture for brain cell parameter measurement.
Main Results:
- Improved segmentation of biological and medical data, yielding better object boundaries and sharper edges.
- Successful processing of data previously intractable with other methods.
- Demonstrated potential for enhanced diagnostic and cellular analysis.
Conclusions:
- The proposed methods, including GrabCut, fuzzy logic, and deep learning, significantly advance image segmentation in medical and biological fields.
- These techniques offer improved accuracy and detail for diagnostic and research applications.
- Further extension to other medical and biological domains is promising.
