U-NET Deep Learning Model on Highway Driving Dataset

Check out the Jupyter Notebook on Google Colab Here:

https://colab.research.google.com/drive/1MC0A4Age2KImRPsniXa9IGpQsvnR-h1t?usp=sharing

A practice in Convolutional Neural Networks (CNN) and attempt at image recognition through pixel classification. This was a project from my Deep Learning course.

Based on the Xception style U-Net. This CNN trained on images trimmed to 720×720 with (RGB) pixels categorized into integers (0-9). Trained largely on Categorical Cross Entropy produced significant results.

Training Model: