Inspiration
Computing is a powerful tool that is currently underused in the medical field today. One of the members on our team, Jenny, is in the Neuroscience Lab at TJ. She wanted to be a doctor growing up, but found a passion for programming in high school. However, she now wants to apply computer science to biology research to find non-invasive methods for diagnoses and treatments of diseases, including cancer.
How it works
A Web application takes a set of DICOM files of horizontal brain slices captured by an MRI and runs a Mathematica function through it. Using the Mathematica function, the tumor is detected using a series of image processing techniques that are able to distinguish cancerous masses by applying distinct morphological, eccentricity, and contrast features typically presented by tumors in MRIs.
Challenges I ran into
- Mathematica running as a standalone executable and on a Web app successfully
- Incorrect detection of tumors (detecting bones as tumors)
Accomplishments that I'm proud of
- Successful detection and visualization of tumors
- 3D representation of MRI images (brain modeling)
- Using front end to execute proprietary Mathematica program on Web application
What I learned
Wolfram Alpha is powerful, but it is not a very mobile language.
What's next for Automated Detection of Brain Tumors from MRI Images
- Clear server of clutter
- Improve detection of unclear, less contrasted MRI images
- Improve detection of tumors with atypical shapes
Log in or sign up for Devpost to join the conversation.