Command Palette
Search for a command to run...
Lo Ka Man
Abstract
Motion detection is a fundamental but challenging task for autonomousdriving. In particular scenes like highway, remote objects have to be paidextra attention for better controlling decision. Aiming at distant vehicles, wetrain a neural network model to classify the motion status using optical flowfield information as the input. The experiments result in high accuracy,showing that our idea is viable and promising. The trained model also achievesan acceptable performance for nearby vehicles. Our work is implemented inPyTorch. Open tools including nuScenes, FastFlowNet and RAFT are used.Visualization videos are available athttps://www.youtube.com/playlist?list=PLVVrWgq4OrlBnRebmkGZO1iDHEksMHKGk .
Code Repositories
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| motion-detection-on-nuscenes | Raft (Kitti) | F1 (%): 89.5 |
| motion-detection-on-nuscenes | FastFlowNet (Kitti) | F1 (%): 92.9 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.