This project predicts the presence of melanoma, nevus, or seborrheic keratosis. It also determines keratinocytic vs melanocytic disease based on pictures of a lesion.
First, you can check out my code here. If you are looking for further instructions, specifics, or the file with the data I used for this paragraph, you can visit this link and follow the recommendations and instructions on how this all works.
- First, you need a Jupyter Notebook. I do all my code through Anaconda.
- Next, you need a GPU or a CPU. I highly recommend a GPU; if you're following along with the Udacity Nanodegree there should be one available.
- Next, you're going to need the files. Go to the recommendations and instructions I attached in the "What You Need to Get Started" Section; I would use 7zip to download the skin diagnosis datasets, as explained by @udacity.
- You also need PyTorch and its corresponding libraries. You can check out some installation instructions on the PyTorch website
- A portion of the project uses a keratinocyte/melanocyte binary classification at the end. I reorganized the files to combine two of them and left the third as its own
@jsreddy3, or Jaiden Reddy! If you have questions you can email me at jaidenreddy@gmail.com