CNNs based analysis to predict diagnosis in Optical Coherence Tomography (OCT) scans

Project Structure

This project had to be divided into 2 different jupyter notebooks. Benchmark.ipynb and InceptionV3.ipynb

Benchmark.ipynb contains the benchmark model and InceptionV3.ipynb contains the optimized transfer learning model.

• Reduced dataset to run this project can be found here: https://drive.google.com/file/d/1alCo05os-s1I1OljXoFjvO1Ik0ikleiG/view?usp=sharing

• Relevant files:

  • proposal.pdf
  • Benchmark.ipynb
  • Benchmark.html
  • InceptionV3.ipynb
  • InceptionV3.html
  • Capstone project report.pdf

Capstone Report

The format of capstone project report follows the template provided by Udacity

The capstone report is part of this repository here is the link: https://github.com/aleks091/machine-learning-capstone/blob/master/Capstone%20project%20report.pdf

Capstone Proposal

Capstone proposal document can be found here:

• https://github.com/aleks091/machine-learning-capstone/blob/master/Capstone%20Proposal.pdf

Capstone proposal review can be found here:

• https://review.udacity.com/#!/reviews/1219487

Software and Libraries

• Keras v2.1.6
• TensorFlow v1.8.0
• CUDA release 9.0, v9.0.176
• Python v3.6.2
• Jupyter Notebook v5.4.0