A Trade Study utilizing deep learning (TensorFlow) in Python and Image Manipulation
Access this project’s GitHub Page:
Description:
- Experimented with different image processing techniques such as changing aperture, image contrast and image composition to determine the specifications optimized for classifying reptiles/amphibians using convolutional neural networks
- Found that our base model, Xception, was improved when brightening image contrast, achieving an ROC score of 0.97
Project Report:
Access this project’s Google Slides presentation:
Erika Fox 