BrainStain.ai 🧠

✅ Module 1: Brain tumor segmentation using U-NET architecture and Sorensen Dice Index metric. Trained on 3D NIfTI data.

✅ Module 2: Multimodal Transfer Learning Approach for Alzheimer’s Disease Stage Classification Using 3D Convolutional Neural Networks. Trained on ADNI data.

For 3D NifTi data:

• Implemented 3D-CNNs:
  • SEResNeXt101 (0.9879 Val AUC)
  • ResNet18 (0.9982 Val AUC)
  • VGG-16 (0.9989 Val AUC)
  • DenseNet121 (0.9902 Val AUC)
• Built a multimodal approach. Modalities of the patient used are:
  • Gender
  • Age
  • Frequency of Visit
  • Patient ID

For 2D data: The 2D data was created by taking the middle slice of the 3D NifTi data.

• Implemented:
  • ViT-B_32 with imagenet21k weights (0.86 Val AUC)

Dependencies (Module 1)

Package	Tested version
tensorflow	2.11.0
keras	2.10.0
opencv_python	4.7.0.68
nibabel	5.0.1
numpy	1.23.3
pandas	1.5.0
matplotlib	3.6.0
scikit_learn	1.2.2
SimpleITK	2.2.1
ipython	8.11.0