An improved brain image classification technique with mining and shape prior segmentation procedure.

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Artificial Intelligence in Medicine

Background:

• Accurate classification of brain tumors from CT scans is crucial for effective treatment planning.
• Existing methods may have limitations in distinguishing between normal, benign, and malignant brain tissues.
• Integrating image features with expert knowledge can potentially improve diagnostic accuracy.

Purpose of the Study:

• To develop an improved brain image classification system using a novel algorithm.
• To classify CT scan brain images into normal, benign, and malignant categories.
• To enhance diagnostic accuracy by combining low-level image features with high-level expert knowledge.

Main Methods:

• The study employed a shape prior segmentation procedure.
• The ImageApriori algorithm with pruned association rules was utilized for classification.
• Low-level image features and high-level expert knowledge were incorporated into the decision process.

Main Results:

• The proposed system achieved 97% sensitivity in classifying brain images.
• Specificity was recorded at 91% for the classification task.
• The overall accuracy of the brain image classification system reached 98.5%.

Conclusions:

• The developed ImageApriori algorithm offers a highly accurate method for brain image classification.
• The system effectively distinguishes between normal, benign, and malignant brain conditions.
• This approach is expected to assist physicians in making more efficient and informed diagnostic decisions.