Human Life Saviour's

Our final year project is about "Genetic subtype of MGMT prediction using MRI." Being an engineer or developer for a humanitarian cause by choosing medical imaging to save people's life was the goal of this project and to be honest, it has been more of a challenge too. Our aim is to design a model that can efficiently and accurately predict the MGMT status of a gene that is either methylated or unmethylated. In the FYP-I, we have a dataset of almost 5 to 5.50 lac MRI scans in which we have 4 types of MRI Scans (t1w, t1wcE, t2w and Flair). At the very first stage, we filter the MRI images by applying multiple filters such as Gaussian noise and mean mode filter and then we train and test our model by implementing autoencoders and sparse autoencoder neural networks only on t1w type MRI scans to predict the MGMT status. Later in FYP-II, we will apply and implement all types of MRI scans and try to achieve high accuracy as possible. The final product is a web-based GUI, that takes the MRI scans of humans and it predicts the MGMT status as either methylated or unmethylated.

Ibrar Babar and Arslan Haider are on my team. We are classmates and roommates in dorm as well. We divided equally the work, and each person is working on their respective tasks.