MULTIPLE TYPES OF CANCER CLASSIFICATION USING CT or MRI IMAGES BASED ON LEARNING WITHOUT FORGETTING POWERED MOBILENETV2 MODELS

This study presents a novel approach for the classification of multiple types of cancer using CT/MRI images, employing state-of-the-art learning methodologies without forgetting. The research leverages MobileNetV2 models to achieve robust and efficient classification performance. The utilization of convolutional neural networks, particularly MobileNetV2, showcases the potential to enhance accuracy and computational efficiency in cancer classification tasks. The proposed methodology focuses on overcoming the challenge of forgetting during sequential learning, ensuring that the model retains information from previously encountered classes while adapting to new ones. The experimental evaluation involves the analysis of CT/MRI images, demonstrating the effectiveness of the proposed approach in accurately classifying diverse cancer types. The integration of MobileNetV2 models contributes to the scalability and adaptability of the classification system. Results indicate promising outcomes in terms of classification accuracy and computational efficiency, establishing the proposed method as a valuable tool for the accurate and efficient categorization of multiple cancer types in medical imaging.

Keywords: Image processing, Multi type cancers dataset, Deep learning, mobileNetv2.