Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

MRI: stability of three supervised segmentation techniques

L P Clarke¹, R P Velthuizen, S Phuphanich

¹Center for Engineering and Medical Image Analysis (CEMIA), College of Engineering, University of South Florida, Tampa 33612.

Magnetic Resonance Imaging

|January 1, 1993

Summary

This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

An Assessment of Imaging Informatics for Precision Medicine in Cancer.

Yearbook of medical informatics·2017

Same author

Clinical outcomes of melanoma brain metastases treated with stereotactic radiosurgery and anti-PD-1 therapy, anti-CTLA-4 therapy, BRAF/MEK inhibitors, BRAF inhibitor, or conventional chemotherapy.

Annals of oncology : official journal of the European Society for Medical Oncology·2016

Same author

A phase II study of bevacizumab and erlotinib after radiation and temozolomide in MGMT unmethylated GBM patients.

Journal of neuro-oncology·2015

Same author

Phase I trial of sorafenib in patients with recurrent or progressive malignant glioma.

Neuro-oncology·2011

Same author

The Reference Image Database to Evaluate Response to therapy in lung cancer (RIDER) project: a resource for the development of change-analysis software.

Clinical pharmacology and therapeutics·2008

Same author

NCI image archive management workshop: a preliminary report.

Academic radiology·2001

Same journal

Incremental diagnostic value of microstructural time-dependent diffusion MRI in differentiating PCNSL from glioblastoma over conventional MRI.

Magnetic resonance imaging·2026

Same journal

Enhanced motion compensation for free-breathing dynamic contrast-enhanced MRI with GROG-facilitated bunch phase encoding and Golden angle radial sampling.

Magnetic resonance imaging·2026

Same journal

The allegory of the cave: 10 years of AI shadows in radiology.

Magnetic resonance imaging·2026

Same journal

Conversion of 3 T liver, spleen, pancreas, and kidney R2* measurements to 1.5 T R2* equivalents: Validation of a theoretical framework.

Magnetic resonance imaging·2026

Same journal

Cine-derived mitral annular relaxation velocity for detection of preclinical left ventricular diastolic dysfunction.

Magnetic resonance imaging·2026

Same journal

Bone marrow fat fraction and R2* in sickle cell disease: Associations with hemolysis, iron metabolism, and disease severity.

Magnetic resonance imaging·2026

See all related articles

The k-nearest neighbors (k-NN) method offers the most stable segmentation for multispectral MRI data, outperforming maximum likelihood (MLM) and artificial neural nets (ANN) in classifying brain tumors.

Area of Science:

Medical Imaging
Artificial Intelligence
Pattern Recognition

Background:

Multispectral MRI data segmentation is crucial for accurate medical diagnosis.
Supervised pattern recognition techniques offer potential for automated segmentation.
Evaluating the stability and speed of these methods is essential for clinical application.

Purpose of the Study:

To compare the performance of three supervised segmentation methods: maximum likelihood (MLM), k-nearest neighbors (k-NN), and artificial neural networks (ANN).
To assess segmentation performance based on execution speed and stability across different data selection criteria (ROI, interslice, interpatient).
To determine the most effective method for segmenting multispectral MRI data in clinical scenarios, including glioma patients.

Main Methods:

Related Experiment Videos

Applied supervised segmentation using MLM, k-NN, and ANN algorithms.
Evaluated segmentation speed and stability using region of interest (ROI) selection, interslice, and interpatient classifications.
Utilized multispectral MRI datasets from normal volunteers and patients with gliomas (with and without contrast material).

Main Results:

Maximum likelihood method (MLM) exhibited the fastest execution times but the lowest stability.
K-nearest neighbors (k-NN) demonstrated superior stability in training data selection.
Across all evaluation measures, k-NN yielded the best segmentation results for multispectral MRI data.

Conclusions:

K-nearest neighbors (k-NN) is the most stable and effective supervised method for segmenting multispectral MRI data.
The findings support the use of k-NN for improved accuracy in brain tumor segmentation and analysis.
Further research can explore k-NN's adaptability to diverse neuroimaging datasets and clinical applications.