CNN-Based Brain Tumor Detection Model Using Local Binary Pattern and Multilayered SVM Classifier.

Area of Science:

• Medical Imaging and Artificial Intelligence
• Computational Neuroscience
• Machine Learning in Healthcare

Background:

• Accurate brain tumor detection and classification are crucial for effective patient treatment and require efficient, automated systems to reduce manual workload.
• Existing methods may lack the precision and speed needed for timely diagnosis, highlighting the need for advanced computational approaches.

Purpose of the Study:

• To develop and evaluate an autonomous model for brain tumor segmentation and detection using advanced deep learning techniques.
• To integrate Local Binary Pattern (LBP) and Multilayered Support Vector Machine (ML-SVM) with Convolutional Neural Networks (CNN) for enhanced performance.

Main Methods:

• Image preprocessing involved filtering and intensity normalization, followed by patch extraction.
• Feature extraction utilized grayscale conversion and Local Binary Pattern (LBP) on RGB images.
• A Convolutional Neural Network (CNN) was employed for feature extraction, and a Multilayered Support Vector Machine (ML-SVM) for object detection and classification.

Main Results:

• The proposed model demonstrated effectiveness in determining the presence or absence of brain tumors.
• Performance was evaluated against existing methods using metrics including Dice Similarity Coefficient (DSC), Jaccard Similarity Index (JSI), sensitivity, accuracy, specificity, and precision.

Conclusions:

• The developed autonomous system offers a promising approach for brain tumor diagnosis, integrating CNN and ML-SVM for robust detection and segmentation.
• This intelligent system has the potential to significantly aid physicians by automating complex diagnostic tasks, improving efficiency and accuracy in clinical practice.