Lyft 3D Object Detection for Autonomous Vehicles
There are three main approaches to be implemented:
- A simple approach using U-net (done)
- Voxelnet
- BEV fusion one of the SOTA and in the top 10 in the Nuscenes benchmark
Acknowledgement and references: Was trained on Kaggle (special thanks) Much of the data understanding and many function are reused from https://www.kaggle.com/code/gzuidhof/reference-model/notebook which is released under the Apache 2.0 open source license