Enhancing Medical Image Classification through Transfer Learning and CLAHE Optimization

Area of Science:

• Medical Imaging
• Computer Vision
• Machine Learning

Background:

• Medical image classification is crucial for diagnosis.
• Transfer learning offers a promising approach for medical image analysis.
• Image preprocessing techniques can enhance model performance.

Purpose of the Study:

• To evaluate the impact of Contrast Limited Adaptive Histogram Equalization (CLAHE) on brain image classification.
• To assess the combined effect of CLAHE and data augmentation in transfer learning models.
• To improve the accuracy and reliability of automated brain image analysis.

Main Methods:

• Four experimental setups were designed: normal images (with/without data augmentation) and CLAHE-processed images (with/without data augmentation).
• Transfer learning models were employed for classification tasks.
• Performance metrics including precision, recall, F1-score, and accuracy were analyzed.

Main Results:

• CLAHE combined with data augmentation yielded superior classification results.
• This optimal setup achieved a precision of 0.90, recall of 0.87, F1-score of 0.89, and accuracy of 0.86.
• The proposed method significantly outperformed other tested configurations.

Conclusions:

• CLAHE optimization is highly effective for enhancing transfer learning in medical image classification.
• The synergistic effect of CLAHE and data augmentation leads to substantial improvements in model performance.
• This study underscores the value of advanced image processing techniques for robust brain image classification.