NanoScan: A Novel Tool For Detecting Hairline Fractures

This adaptive tool that diagnoses hairline fractures, small cracks in the bone that are undetectable to human eye, is a novel ML approach to diagnosing these fractures that become worse if untreated.

Comment 1

Inspiration

After an ankle injury that was said to be a "sprain" but actually turned out to be a hairline fracture, I wanted to create something that is certain about diagnosing hairline fractures

What it does

Takes in images of bone x-rays and determines whether or not the bone is fractured or normal

How we built it

Used Python in Colab, a binary classification system for the two fractured vs. not fractured classes, used the sigmoid function with the binary classification

Challenges we ran into

The model was quite slow the first time around but I made some adjustments

Accomplishments that we're proud of

Making a model that connects to my own personal story and that could potentially be used in the medical field!

What we learned

I learned more about neural networks, as before I had focused on KNN

What's next for NanoScan: A Novel Tool For Detecting Hairline Fractures

Use in the medical field as a product that detects hairline fractures :)

Built With

  • colab
  • matplotlib
  • python
  • tensorflow
Share this project:

Updates