This repository contains visual odometry baselines that you can evaluate on several datasets. Currently provided methods: -LS-VO -Resnet50
Currently supported datasets: -KITTI -TUM -DISCOMAN
- git clone --recursive https://github.com/vul-samsung/odometry
- conda env update -f conda.yml
- python download_weights.py
- Download datasets from official websites
- Prepare dataset for training with scriptsprepare_dataset.py
- Configure paths in env.py
Start training with scripts/train.py