This repo aims to provide flexible and reproducible 3D object detection on KITTI dataset. I expect scripts starting from the current directory, and treat ./visualDet3D as a package that one could modify and test directly instead of a library. Several useful scripts are provided in the main directory for easy usage.
I believe that visual tasks are interconnected. The package uses registry to register datasets, models, processing functions and more, allowing easy inserting of new tasks/models while not interfere with the existing ones.
- SOTA Performance State of the art result on 3D object detection.
- Modular Design Modular design for dataset, network and running pipelines.
- Support Various Task Compatible with the training and testing of mono/stereo 3D detection and depth prediction.
- Distributed & Single GPU Support training with multiple GPUs.
- Installation-Free Setup The setup process only build operations and does not require installation to keep the environment clean.
- Global Path-based IMDB Do not need data placed inside the folder, convienient for managing data and code separately.
Reference: this repo borrows codes and ideas from retinanet, mmdetection, M3D-RPN, DORN, EdgeNets, det3
In this repo, the folders named as - "config", "objectDet3D" and "scripts" are the ones created by me.