Project description:
Early detection and classification of brain tumors is an important research domain in the field of medical imaging and accordingly helps in selecting the most convenient treatment method to save patients life therefore.Our project aims to help medical practitioners speed up the classification of MRI scans for brain tumors.
You'll need the following libraries for the provided imports:
- NumPy for numerical operations:
import numpy as np
- Pandas for data manipulation and analysis:
import pandas as pd
- Matplotlib for visualization:
import matplotlib.pyplot as plt
andfrom matplotlib import cm
- Torch and Torchvision for deep learning:
import torch
andimport torchvision
- Torch.utils.data for handling datasets:
from torch.utils.data import Dataset
- Torch.optim for optimization algorithms:
import torch.optim as optim
- Torchvision.transforms for data transformations:
from torchvision.transforms import ToTensor, Compose, Normalize
- Torchvision.datasets for accessing datasets:
from torchvision.datasets import MNIST
- Torch.nn.functional and Torch.nn for neural network operations:
import torch.nn.functional as F
andimport torch.nn as nn
- Math for mathematical operations:
from math import ceil
link to weights:
DenseNet: https://drive.google.com/drive/folders/1SjDfoUmSJUFgIBRG02Atkv8_vnq0WVOs?usp=drive_link
ResNet: https://drive.google.com/drive/folders/15ylRamh0m1PQG_eX27bdEoyR-kNTL0GV?usp=drive_link
EfficientNet: https://drive.google.com/drive/folders/1Ig_ces5s4Xcuezr2BU4TXhbuT94dy_OZ
VGG: https://drive.google.com/drive/folders/1BGJUxurT9y4FvaI5r6EvPYDPZ76PYMGV
Alexnet : http://login2.gpucluster.sutd.edu.sg/hub/user-redirect/lab/tree/AlexNet_weights
Alexnet_transfer_learning : https://sutdapac-my.sharepoint.com/:f:/g/personal/vinny_koh_mymail_sutd_edu_sg/ElglhwrA2KtGm8Jdj5RIBGEByTYjj2tLddHbnucN7NuUTg?e=47BWYG
CNN weights : http://login2.gpucluster.sutd.edu.sg/hub/user-redirect/lab/tree/CNN_weights
(unfortunately for AlexNet and CNN , we had difficulties moving it to one drive due to its large size - 12GB and 42 gB , thus it could only be accessed from the gpu cluster where we trained it - however if unable to access GPU cluster , the tester codes to prove reproducibility has been run already , we are sorry for that!)
link to data: https://www.kaggle.com/datasets/masoudnickparvar/brain-tumor-mri-dataset
link to report: https://docs.google.com/document/d/1WhZwx9Ngb4_aex9ah9Wu2sTVc3zfQ8Kkieg4d0xD0Vw/edit?usp=sharing