# lerojd

**Repository Path**: fightingand/lerojd

## Basic Information

- **Project Name**: lerojd
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-10-01
- **Last Updated**: 2024-10-01

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Towards Efficient 3D Object Detection with Knowledge Distillation

---

This repository contains the official implementation of [LEROjD: Lidar Extended Radar-Only Object Detection](), Accepted for publication at ECCV 2024


## Changelog
[2024-08-] Initial release

## Introduction
Our code is based on [SparseKD](https://github.com/CVMI-Lab/SparseKD) which itself is based on [OpenPCDet v0.5.2](https://github.com/open-mmlab/OpenPCDet/tree/v0.5.2).
Some changes of [OpenPCDet v0.6.0](https://github.com/open-mmlab/OpenPCDet/tree/master), like DSVT, are integrated in this codebase.

We only supply code for execution on the [View of Delft]((https://github.com/tudelft-iv/view-of-delft-dataset)). The code should be applicable to other datasets with minor changes.

Due to the limited availability of evaluation on the [View of Delft](https://github.com/tudelft-iv/view-of-delft-dataset) Test dataset we use a different split of the dataset for training and evaluation.
This modified split is available in the [data](data/view_of_delft_PUBLIC/lidar/ImageSets) folder.

## Installation

Please refer to [INSTALL.md](docs/INSTALL.md) for the installation.

## Getting Started

Please refer to [GETTING_STARTED.md](docs/GETTING_STARTED.md) to learn more usage of OpenPCDet.

## License
This codebase is released under the [Apache 2.0 license](LICENSE).

## Acknowledgement
Our code is heavily based on [SparseKD](https://github.com/CVMI-Lab/SparseKD) and [OpenPCDet](https://github.com/open-mmlab/OpenPCDet). 
Thanks to the authors of SparseKD as well as the OpenPCDet Development Team for their awesome codebase.