ECIENet

Introduction

This is the PyTorch implementation of our paper "Monocular 3D object detection based on context information enhancement and deep guidance".

Overview

• Installation
• Getting Started

Installation

We test this repository on one Nvidia 2080Ti GPU.

Ubuntu20.04

python=3.6.13

torch=1.8.1

Getting Started

Dataset Preparation

KITTI

• Please download the official KITTI 3D object detection dataset and organize the downloaded files as follows:

│── KITTI
│   │── ImageSets
│   │   ├──train.txt & val.txt & trainval.txt & test.txt
│   │── training
│   │   ├──calib & image_2 & label_2 & depth_dense
│   │── testing
│   │   ├──calib & image_2

• The dense depth files at: Google Drive

Waymo

You can follow the Blog to convert the Waymo Dataset.

Training & Testing

• Test our paper's result

upload the add_DL_lapulasi_lr09.zip to KITTI benchmark

• Test and evaluate the pretrained models

python tools/test.py 

• Train a model

python tools/train_val.py