부산대학교 도서관

In the recent scenario, the move from machine learning to deep learning has played a significant role in various applications such as, healthcare, automation, robotics, etc. Considerign the healthcare, diagnosing diseases, like brain tumor, brest cancer, etc. using Magnetic Resonance Imaging (MRI) images, is a complex process because the images contain various cell/image pixel divisions to identify the condition of the disease. Based on the complexity of the disease, identifying and diagnosing the tumour becomes a challenging task as it may lead to an increase in the medical survey on tumour disease based on the number of patients. In this paper, a Deep Learning (DL) - Convolutional Neural Networks (CNN) based approach is proposed with the integration of data augmentation and image processing. The proposed approach provides classification of MRI images to diagnose the tumour location on the dense areas based on small tumour cells. The effectiveness of the proposed approach is analyzed based on the data accuracy as it achieves 96% with low complexity rate compared to the existing models such as, CNN and Artificial Neural Networks (ANN) models.