Breast Cancer Classification through Meta-Learning Ensemble Technique Using Convolution Neural Networks.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Computational Pathology

Background:

• Accurate classification of breast lesions in ultrasound images is critical for early breast cancer detection and treatment.
• Traditional machine learning and deep learning models struggle with the complexity and diversity of breast ultrasound images (BUSI).

Purpose of the Study:

• To develop an efficient and accurate breast cancer classification model using meta-learning and multiple convolutional neural networks (CNNs).
• To improve the classification of benign versus malignant breast lesions in the BUSI dataset.

Main Methods:

• Utilized a meta-learning ensemble technique combined with transfer learning (Inception, ResNet50, DenseNet121) and data augmentation.
• Trained and evaluated multiple CNN architectures on the BUSI dataset, optimizing learning with meta-learning algorithms.
• Employed ensemble learning to combine outputs from diverse CNNs for enhanced classification accuracy.

Main Results:

• The proposed model demonstrated high effectiveness and accuracy in classifying breast ultrasound images.
• Performance metrics including accuracy, precision, recall, and F1 score were evaluated.
• Results showed superior performance compared to existing state-of-the-art approaches.

Conclusions:

• The meta-learning ensemble approach significantly enhances breast cancer classification accuracy from ultrasound images.
• This model offers a promising tool for improving early breast cancer diagnosis and patient outcomes.
• Further validation and comparison with state-of-the-art methods confirm the model's efficacy.