jstol / road-estimation

Investigating machine learning algorithms for identifying roads within dashcam photos

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

road-estimation

Analyzing the effectiveness of various machine mearning algorithms in the segmentation/identification of roads within images from vehicle-mounted cameras.

We are making use of the KITTI Vision Benchmark Suite to evaluate models.

Libraries used:
Examples:

Original Photo: original-photo/um_road_000042 Road Prediction - Brighter means higher predicted probability of being road: encoded/um_road_000042 Road Prediction Overlayed on Original Photo: encoded-overlay/um_road_000042

Original Photo: original-photo/um_road_000082 Road Prediction - Brighter means higher predicted probability of being road: encoded/um_road_000082 Road Prediction Overlayed on Original Photo: encoded-overlay/um_road_000082

About

Investigating machine learning algorithms for identifying roads within dashcam photos


Languages

Language:Python 74.3%Language:Shell 25.7%