Related Experiment Video
Updated: Oct 19, 2025
06:48
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
9.1K
A Survey of Brain Tumor Segmentation and Classification Algorithms
Erena Siyoum Biratu1, Friedhelm Schwenker2, Yehualashet Megersa Ayano3
1College of Electrical and Mechanical Engineering, Addis Ababa Science and Technology University, Addis Ababa 120611, Ethiopia.
Journal of Imaging
|September 26, 2021
Summary
Automated brain tumor segmentation and classification using Magnetic Resonance Imaging (MRI) offers a safer, non-invasive diagnostic alternative to biopsies. This survey comprehensively reviews region growing, shallow machine learning, and deep learning techniques.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Neuro-oncology
Background:
- Manual segmentation of brain tumors from Magnetic Resonance Imaging (MRI) is labor-intensive and time-consuming.
- Automated brain tumor classification from MRI scans provides a non-invasive diagnostic approach, enhancing patient safety by avoiding biopsies.
- Significant research efforts since the late 1990s have focused on developing automatic methods for brain tumor segmentation and classification.
Purpose of the Study:
- To provide a comprehensive survey of recently proposed, major brain tumor segmentation and classification model techniques.
- To analyze and compare three key approaches: region growing, shallow machine learning, and deep learning.
- To cover technical aspects including strengths, weaknesses, pre/post-processing, feature extraction, datasets, and performance evaluation.
Main Methods:
- Review of established literature on brain tumor segmentation and classification.
- Focus on three primary methodologies: region growing, shallow machine learning, and deep learning.
- Analysis of technical components such as pre-processing, feature extraction, and performance metrics.
Main Results:
- Extensive literature exists on brain tumor segmentation using region growing, traditional machine learning, and deep learning.
- Impressive performance results have been achieved in classifying brain tumors into histological types.
- The survey consolidates information on the strengths and weaknesses of different segmentation and classification approaches.
Conclusions:
- Automated methods for brain tumor segmentation and classification are crucial for efficient and safe cancer diagnosis.
- The survey highlights the advancements and comparative performance of region growing, shallow machine learning, and deep learning techniques.
- Understanding these methods aids in selecting appropriate models for clinical application and future research.
