jeffrey@penn:~/projects/cone-segmentation$ cat README.md
Binary segmentation for cone detection
Real-time traffic-cone segmentation for FSAE Driverless.
A lightweight binary segmentation model with a simple encoder-decoder architecture that detects traffic cones in real time for autonomous driving, specifically the FSAE Driverless competition.
It reaches high accuracy at low latency, which makes it small and fast enough to deploy on edge devices in the car.