An Image Processing-based and Deep Learning Model to Classify Brain Cancer

In recent years, the prevalence of cancer has increased significantly around the world. Cancer is considered one of the most dangerous diseases in humans. Cancer screening devices, such as Magnetic Resonance Imaging (MRI), X-ray imaging, ultrasound imaging, and others, play an important role in its early detection. This study aims to facilitate cancer tumor detection on mobile phones with high accuracy in a short period of time using deep learning techniques. A brain tumor dataset was used, consisting of 4,489 images and 14 classified types, and experiments were carried out using ResNet 12, DenseNet, YOLOv8, and MobileNet to evaluate them in terms of accuracy, speed, and model size. ResNet12, DenseNet, YOLOv8, and MobileNet results indicated satisfactory accuracy ranging from 88% to 97.3%. YOLOv8 was the most suitable model, as its fastest inference time of 1.8 ms, preprocessing time of 0.1 ms, highest accuracy of 97.3%, and compact model size make it ideal for real-time mobile applications.