Commit a9980a55 authored by tl2120's avatar tl2120
Browse files

Replace README.md

parent 1fb14f0b
# Attack and Defences on LiDAR-based Perception Documentation
#### Release 0.0.1
###### 02 September 2021
<div style="page-break-after: always;"></div>
*Attack and Defences Repository* explores different LiDAR-based perception module, namely PIXOR, PointRCNN and PointPillar - in the form of Baidu Apollo v6.0
<div style="page-break-after: always;"></div>
# Table of Contents
* PIXOR
* Dependencies
* File Structure
* Getting Started
* Download Kitti Dataset
* Run PIXOR Perception Module
* View Results
* PointRCNN
* Dependencies
* Baidu Apollo
* Docker load
<div style="page-break-after: always;"></div>
# PIXOR
##### Dependencies:
- Python 3.5(3.6)
- Pytorch (Follow Official Installation Guideline)
- Tensorflow (see their website)
- Numpy, MatplotLib, OpenCV3
- PyKitti (for running on KITTI raw dataset)
- gcc
Note: All requirements are preloaded into pixorenv
#### File Structure
```
attack_and_defences...
├── PIXOR
│ ├── Kitti_Dataset
| | ├── raw_data_downloader.sh
| | ├── 2011_09_26
| | ...
| ├── PIXOR
| | ├── srcs
| | ... ├── run_kitti.py
| ├── pixorenv
│ │ ├── bin
| | | ├── activate
| | | ....
```
##### Getting Started
Ensure that virtualenv is installed
$ pip install virtualenv
>
Activate PIXOR virtual environment
$ cd PIXOR
$ source pixorenv/bin/activate
##### Download Kitti Dataset
$ cd Kitti_Dataset
$ ./raw_data_downloader.sh
##### Run PIXOR Perception Module
$ cd PIXOR
$ cd srcs
$ python run_kitti.py
##### View Results
Depending on the video that was processed, typically located in PIXOR/Kitti_Dataset/2011_09_26/ ...
<div style="page-break-after: always;"></div>
# PointRCNN
##### Dependencies:
All the codes are tested in the following environment:
* Linux (tested on Ubuntu 14.04/16.04)
* Python 3.6+
* PyTorch 1.0
Note: All requirements are preloaded into conda environment
#### File Structure
```
attack_and_defences...
├── PointRCNN
│ ├── conda
| | ├── bin
| | | ├── activate
| | ...
| ├── PointRCNN
| | ├── tools
| | ... ├── eval_rcnn.py
| | ├── PointRCNN.pth
| | ├── output
```
##### Getting Started
Ensure that Anaconda or Miniconda is installed
Activate conda virtual environment
$ cd PointRCNN
$ ./conda/bin/activate
##### Run PIXOR Perception Module
$ cd PointRCNN
$ cd tools
$ python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt <Path_To_PointRCNN.pth> --batch_size 1 --eval_mode rcnn --set RPN.LOC_XZ_FINE False
##### View Results
Depending on the video that was processed, typically located in PointRCNN/output/ ...
# Apollo Baidu Docker Image
*Requires Docker*
#### File Structure
```
attack_and_defences...
├── apollo
│ ├── modules
| | ├── pcpc
| | | ├── BUILD
| | | ├── talker.cc
| | | ├── adv_loss.py
| | | ├── logs
| | | | ├── cls_score.txt
| | | | ├── bbox_pred.txt
| | ...
| ├── bazel-bin
| | ├── modules
| | ... ├── pcpc
| | | | ├── talker
| | ├── output
```
##### Run build
From root directory /apollo
$ ./apollo.sh build pcpc
##### Run talker
*To update cls_score.txt and bbox_pred.txt*
$ ./bazel-bin/modules/pcpc/talker
$ cd tools
#### Run adv_loss.py
$ python modules/pcpc/adv_loss.py
\ No newline at end of file
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment