Brain Tumour Classification using Deep Learning

A new age of medicine

Image provided by the author of a brain MRI of a glioma.

Brain Tumours

Brain tumours are every patient's worst nightmare — headaches, dizziness, blurred vision, seizures, impaired judgement, and emotional volatility can all point to a brain tumour, but they can also point to nothing. Before the hypochondriac jumps to conclusions or the busy patient declares “it’s nothing,” physicians generally follow three steps to diagnose brain tumours:

1. A Neurological Exam
2. Neuro Scan (MRI, CT, Angiogram, etc)
3. A biopsy

Brain tumours are a group of abnormal cells that grow together in the brain to form a mass. These tumours are often fatal. However, there are many different types of brain tumours so physicians sort them into different categories: benign or malignant, primary or metastatic, operable or inoperable. It is important to note that different categories will often have overlapping characteristics.

Brain tumours are named after the cells or area that they originate from. The dataset used in this classification includes:

• Gliomas
• Meningiomas
• Pituitary tumours

Gliomas are named after the cells that they originate from: glial cells. Glial cells are…