Auto-detect coverage bounding boxes for brain MRI images - requirements

[madtitan02]

Hello, I'm Utkarsh, and I am pursuing biomedical engineering at IIT BHU.
@zeydabadi I had some doubts while writing proposal for project -6, which I have listed below

• In the project description, the need for the detection of anatomical coverage, but it is not mentioned what specific protocols need to be included .so while building a demo project for my proposal, it would be helpful if you could provide some information related and what are the variations in protocols for different types of brain MRI scans. Or some reference to some documentation for these protocol constraints would be useful

And i have proposed the following workflow implementation for the project .I could interpret as much as I could from the summary of the project and my past experience.Please go through it once and any suggestion would definitely help to improve it

• preprocessing the DICOM to a suitable file format for our model and normalizing the data
• Then further anatomical constraints detection.
• Then, the determination of bounding boxes based on the detection in previous step and could use deep learning models in this step to precise the data.
• Then comparison of bounding box with the expected and calculation of the compliance score based on the extent of overlap
• validation
• Testing

Above is the basic workflow i have decided and if i got anything wrong do let me know or Suggestions would be benefitial