An Efficient Method for Brain Tumor Detection Using Texture Features and SVM Classifier in MR Images

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Accurate detection and classification of brain abnormalities in Magnetic Resonance (MR) images are critical for medical diagnosis.
• Image noise significantly hinders the retrieval of diagnostic information from MR scans.
• Existing methods require robust techniques for effective feature extraction and classification.

Purpose of the Study:

• To develop and evaluate a comprehensive system for brain tumor detection and classification using MR images.
• To address the challenge of noise in medical imaging through advanced denoising techniques.
• To compare the performance of different feature extraction and classification algorithms.

Main Methods:

• Image denoising was performed using the PURE-LET transform to preserve image quality.
• Feature extraction involved a combination of Modified Multi-Texton Histogram (MMTH) and Multi-Texton Microstructure Descriptor (MTMD), alongside Gray Level Co-occurrence Matrix (GLCM) and Gray Level Run Length Matrix (GLRLM).
• Classification was conducted using Support Vector Machine (SVM), K Nearest Neighbors (KNN), and Extreme Learning Machine (ELM).

Main Results:

• The performance of various feature extraction and classification combinations was evaluated using sensitivity, specificity, and accuracy metrics.
• The proposed system demonstrated effective noise reduction while preserving diagnostic properties.
• The combination of MMTH and MTMD feature extraction with SVM classification yielded the highest accuracy.

Conclusions:

• The integrated system, particularly the MMTH and MTMD feature extraction combined with SVM classification, achieved a high accuracy of 95% for brain tumor classification.
• The PURE-LET transform proved effective in denoising MR images without compromising diagnostic information.
• This approach offers a promising solution for improving the accuracy and reliability of automated brain tumor detection and classification systems.