# Complex-YOLO

**Repository Path**: yajun-wang11/Complex-YOLO

## Basic Information

- **Project Name**: Complex-YOLO
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 1
- **Created**: 2022-04-18
- **Last Updated**: 2022-04-18

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Complex-YOLO
Complex-YOLO: Real-time 3D Object Detection on Point Clouds   pytorch

# Introduction
This is an unofficial implementation of Complex-YOLO: Real-time 3D Object Detection on Point Clouds in pytorch. A large part of this project is based on the work here:https://github.com/marvis/pytorch-yolo2


Point Cloud Preprocessing is based on:https://github.com/skyhehe123/VoxelNet-pytorch
                                      https://github.com/dongwoohhh/MV3D-Pytorch


# Data Preparation

Download the 3D KITTI detection dataset.

Camera calibration matrices of object data set (16 MB)

Training labels of object data set (5 MB)

Velodyne point clouds (29 GB)


# Train

python3 main.py

trained model(using DarkNet) download link :https://pan.baidu.com/s/1yeU3Q-Oyozv7qFqrVhzj0A

# Result

![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/1.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/2.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/3.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/4.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/5.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/6.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/7.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/8.png)
![](https://github.com/AI-liu/Complex-YOLO/blob/master/results/9.png)