Related Experiment Videos
Tumour volume determination from MR images by morphological segmentation
P Gibbs1, D L Buckley, S J Blackband
1Department of Medical Physics, Royal Hull Hospitals NHS Trust, UK.
Physics in Medicine and Biology
|November 1, 1996
Summary
Automated MRI segmentation using morphological edge detection and region growing offers accurate tumour volume measurement. This method is faster and less subjective than manual segmentation for intracerebral glioma.
Area of Science:
- Medical Imaging
- Radiology
- Computational Anatomy
Background:
- Accurate tumour volume measurement from MRI is crucial for treatment planning.
- Manual segmentation is subjective and time-consuming.
- Objective and automated methods are needed for reliable image segmentation.
Purpose of the Study:
- To implement and test an automated segmentation procedure for tumour volume measurement.
- To compare the performance of morphological segmentation with traditional data thresholding.
- To evaluate the efficiency and accuracy of the new segmentation method.
Main Methods:
- A novel segmentation procedure based on morphological edge detection and region growing was developed.
- The method was tested on a phantom with known volume.
- Comparisons were made with data thresholding on patient MR images of intracerebral glioma.
Main Results:
- The morphological segmentation procedure yielded results comparable to traditional data thresholding.
- The automated method demonstrated similar accuracy in tumour volume determination.
- The implemented procedure proved to be faster and automated.
Conclusions:
- Morphological segmentation offers an automated and efficient alternative for tumour volume measurement in MRI.
- This technique reduces operator subjectivity and saves time in clinical practice.
- The method shows promise for accurate segmentation of intracerebral glioma.