top of page
Virtual Tech. Expo
2020
Group 37
Project Overview
Abstract
Our project builds a lane detection and classification system for ADAS application. We first implemented popular nural network algorithms and tested them on public lane detection datasets. Then we did improvement by incorporating self-attention distillation method into ENet, which reached 95.24% accuracy on TuSimple dataset. Finally, we deployed our model into application using Python Flask API.
Poster
Poster
Abstract
Video
Explanatory Video
​Supplementary Video on Expo Poster
bottom of page