Visual Transformers for 3D medical images Classification: use-case neurodegenerative disorders

Title	Visual Transformers for 3D medical images Classification: use-case neurodegenerative disorders
Summary	Using visual transformers for predicting the diagnosis of multiple neurodegenerative brain disorders
Keywords
TimeFrame
References	Attention is all you need: https://arxiv.org/abs/1706.03762 Bert: Pre-training of deep bidirectional transformers for language understanding: https://arxiv.org/abs/1810.04805 Transformers in Vision: A Survey: https://arxiv.org/pdf/2101.01169.pdf ViT-V-Net: https://pythonrepo.com/repo/junyuchen245-ViT-V-Net_for_3D_Image_Registration_Pytorch Our paper, adopting VGG architecture for classification of 3D brain scans: https://link.springer.com/article/10.1007/s00259-021-05483-0
Prerequisites
Author
Supervisor	Stefan Byttner, Kobra Etminani, Amira Soliman
Level	Master
Status	Open

3D PET scans show 3D images of the cell activity in the tissues of the human brain. Having these scans, doctors can use the computer-aided diagnosis of dementia disorders like Alzheimer’s and Parkinson's. Transformers, based on the Bert model, have great success with NLP and are now applied to images under the name of Visual Transformer(ViT). In this thesis, the objective is to apply ViT using brain scans and analyze obtained performance compared to previously adopted state-of-the-art CNN models in terms of accuracy, generated representation, and explainability of decisions taken by ViT.

We had master students last year who worked on adopting 3D CNN for classification purposes. You can check this thesis for example http://urn.kb.se/resolve?urn=urn%3Anbn%3Ase%3Ahh%3Adiva-45370