Inspiration

Healthcare applications for Machine Learning are extremely promising, helping healthcare practitioners in diagnosing serious diseases.

What it does

The application has a web interface, where a user can upload an image of a histopathology slide, and get a predication for if contains being benign or malignant tumor cells.

How I built it

Using Pytorch, we trained our model using the PCam[1][2] slides dataset from https://github.com/basveeling/pcam , and built a flask application and a user interface for uploading the images, the application is deployed to Heroku.

Challenges I ran into

Optimizing the parameters to obtain high accuracy

Accomplishments that I'm proud of

Taking my first steps to experience the already promising machine learning application in healthcare

What I learned

The process of deploying training and optimizing machine learning models.

What's next for Predict Health - Cancer Predication with PyTorch

Optimize for higher accuracy

[1] B. S. Veeling, J. Linmans, J. Winkens, T. Cohen, M. Welling. "Rotation Equivariant CNNs for Digital Pathology". arXiv:1806.03962

A citation of the original Camelyon16 dataset paper is appreciated as well:

[2] Ehteshami Bejnordi et al. Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women With Breast Cancer. JAMA: The Journal of the American Medical Association, 318(22), 2199–2210. doi:jama.2017.14585

Share this project:
×

Updates