Transfer Learning on SAR Imaging

This project was done as part of my summer 2020 internship at the Air Force Research Lab in Rome, New York at the Griffiss Institute. Working with the SAGE in-house R&D team, I was tasked with leveraging transfer learning techniques to perform image classification on spotlight SAR imaging from MSTARS dataset.

Using Python, I implemented a transfer learning model to classify spotlight SAR images with up to 97% test accuracy. Additionally, I performed analysis on the transfer learning approaches to determine the viability of using these methods versus standard machine learning techniques using statistical and hardware measures.