what
WHite-box Adversarial Toolbox (WHAT) is a python library for Deep Learning Security that focuses on realtime white-box attacks.
Installation
pip install whitebox-adversarial-toolbox
Then you can use the cli tool what
to try real-time adversarial attacks.
sage: what [OPTIONS] COMMAND [ARGS]...
The CLI tool for WHite-box Adversarial Toolbox (WHAT).
Options:
--help Show this message and exit.
Commands:
attack Manage Attacks
example Manage Examples
model Manage Deep Learning Models
what.models
Use what model list
to list available models:
Model Model Type Description
----------------------------------------------------------------------------------------------------
[ ] 1 : YOLOv3 ( Darknet ) Object Detection YOLOv3 pretrained on MS COCO dataset.
[ ] 2 : YOLOv3 ( Mobilenet ) Object Detection YOLOv3 pretrained on MS COCO dataset.
[ ] 3 : YOLOv3 Tiny ( Darknet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset.
[ ] 4 : YOLOv3 Tiny ( MobileNet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset.
[ ] 5 : YOLOv4 ( Darknet ) Object Detection YOLOv4 pretrained on MS COCO dataset.
[ ] 6 : YOLOv4 Tiny ( Darknet ) Object Detection YOLOv4 Tiny pretrained on MS COCO dataset.
[ ] 7 : SSD ( MobileNet v1 ) Object Detection SSD pretrained on VOC-2012 dataset.
[ ] 8 : SSD ( MobileNet v2 ) Object Detection SSD pretrained on VOC-2012 dataset.
[ ] 9 : FasterRCNN ( VGG16 ) Object Detection Faster-RCNN pretrained on VOC-2012 dataset.
Use what model download
to download pre-trained models:
Model Model Type Description
----------------------------------------------------------------------------------------------------
[x] 1 : YOLOv3 ( Darknet ) Object Detection YOLOv3 pretrained on MS COCO dataset.
[x] 2 : YOLOv3 ( Mobilenet ) Object Detection YOLOv3 pretrained on MS COCO dataset.
[x] 3 : YOLOv3 Tiny ( Darknet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset.
[x] 4 : YOLOv3 Tiny ( MobileNet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset.
[x] 5 : YOLOv4 ( Darknet ) Object Detection YOLOv4 pretrained on MS COCO dataset.
[x] 6 : YOLOv4 Tiny ( Darknet ) Object Detection YOLOv4 Tiny pretrained on MS COCO dataset.
[x] 7 : SSD ( MobileNet v1 ) Object Detection SSD pretrained on VOC-2012 dataset.
[x] 8 : SSD ( MobileNet v2 ) Object Detection SSD pretrained on VOC-2012 dataset.
[x] 9 : FasterRCNN ( VGG16 ) Object Detection Faster-RCNN pretrained on VOC-2012 dataset.
Please input the model index:
what.attacks
Use what attack list
to list available attacks:
1 : TOG Attack Object Detection
2 : PCB Attack Object Detection
Related Papers:
- Adversarial Objectness Gradient Attacks in Real-time Object Detection Systems.
- A Man-in-the-Middle Attack against Object Detection Systems.
what.examples
Use what example list
to list available examples:
Demo Type Description
--------------------------------------------------------------------------------
1 : Yolov3 Demo Model Inference Yolov3 Object Detection.
2 : Yolov4 Demo Model Inference Yolov4 Object Detection.
3 : FasterRCNN Demo Model Inference FRCNN Object Detection.
4 : MobileNet SSD Demo Model Inference MobileNet SSD Object Detection.
5 : TOG Attack Demo Adversarial Attack Real-time TOG Attack against Yolov3 Tiny.
6 : PCB Attack Demo Adversarial Attack Real-time PCB Attack against Yolov3 Tiny.
Use what example run
to run examples.
Demo Type Description
--------------------------------------------------------------------------------
1 : Yolov3 Demo Model Inference Yolov3 Object Detection.
2 : Yolov4 Demo Model Inference Yolov4 Object Detection.
3 : FasterRCNN Demo Model Inference FRCNN Object Detection.
4 : MobileNet SSD Demo Model Inference MobileNet SSD Object Detection.
5 : TOG Attack Demo Adversarial Attack Real-time TOG Attack against Yolov3 Tiny.
6 : PCB Attack Demo Adversarial Attack Real-time PCB Attack against Yolov3 Tiny.
Please input the example index:
what.utils
This module implements several utility functions.
1r''' 2WHite-box Adversarial Toolbox (WHAT) is a python library for Deep Learning Security that focuses on realtime white-box attacks. 3 4## Installation 5 6``` 7pip install whitebox-adversarial-toolbox 8``` 9 10Then you can use the cli tool `what` to try real-time adversarial attacks. 11 12``` 13sage: what [OPTIONS] COMMAND [ARGS]... 14 15 The CLI tool for WHite-box Adversarial Toolbox (WHAT). 16 17Options: 18 --help Show this message and exit. 19 20Commands: 21 attack Manage Attacks 22 example Manage Examples 23 model Manage Deep Learning Models 24``` 25 26<br /> 27 28## what.models 29 30Use `what model list` to list available models: 31 32``` 33 Model Model Type Description 34---------------------------------------------------------------------------------------------------- 35[ ] 1 : YOLOv3 ( Darknet ) Object Detection YOLOv3 pretrained on MS COCO dataset. 36[ ] 2 : YOLOv3 ( Mobilenet ) Object Detection YOLOv3 pretrained on MS COCO dataset. 37[ ] 3 : YOLOv3 Tiny ( Darknet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset. 38[ ] 4 : YOLOv3 Tiny ( MobileNet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset. 39[ ] 5 : YOLOv4 ( Darknet ) Object Detection YOLOv4 pretrained on MS COCO dataset. 40[ ] 6 : YOLOv4 Tiny ( Darknet ) Object Detection YOLOv4 Tiny pretrained on MS COCO dataset. 41[ ] 7 : SSD ( MobileNet v1 ) Object Detection SSD pretrained on VOC-2012 dataset. 42[ ] 8 : SSD ( MobileNet v2 ) Object Detection SSD pretrained on VOC-2012 dataset. 43[ ] 9 : FasterRCNN ( VGG16 ) Object Detection Faster-RCNN pretrained on VOC-2012 dataset. 44``` 45 46Use `what model download` to download pre-trained models: 47 48``` 49 Model Model Type Description 50---------------------------------------------------------------------------------------------------- 51[x] 1 : YOLOv3 ( Darknet ) Object Detection YOLOv3 pretrained on MS COCO dataset. 52[x] 2 : YOLOv3 ( Mobilenet ) Object Detection YOLOv3 pretrained on MS COCO dataset. 53[x] 3 : YOLOv3 Tiny ( Darknet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset. 54[x] 4 : YOLOv3 Tiny ( MobileNet ) Object Detection YOLOv3 Tiny pretrained on MS COCO dataset. 55[x] 5 : YOLOv4 ( Darknet ) Object Detection YOLOv4 pretrained on MS COCO dataset. 56[x] 6 : YOLOv4 Tiny ( Darknet ) Object Detection YOLOv4 Tiny pretrained on MS COCO dataset. 57[x] 7 : SSD ( MobileNet v1 ) Object Detection SSD pretrained on VOC-2012 dataset. 58[x] 8 : SSD ( MobileNet v2 ) Object Detection SSD pretrained on VOC-2012 dataset. 59[x] 9 : FasterRCNN ( VGG16 ) Object Detection Faster-RCNN pretrained on VOC-2012 dataset. 60 61Please input the model index: 62``` 63 64<br /> 65 66## what.attacks 67 68Use `what attack list` to list available attacks: 69 70``` 711 : TOG Attack Object Detection 722 : PCB Attack Object Detection 73``` 74 75Related Papers: 76 77- [Adversarial Objectness Gradient Attacks in Real-time Object Detection Systems](https://ieeexplore.ieee.org/document/9325397). 78- [A Man-in-the-Middle Attack against Object Detection Systems](https://arxiv.org/abs/2208.07174). 79 80<br /> 81 82## what.examples 83 84Use `what example list` to list available examples: 85 86``` 87 Demo Type Description 88-------------------------------------------------------------------------------- 891 : Yolov3 Demo Model Inference Yolov3 Object Detection. 902 : Yolov4 Demo Model Inference Yolov4 Object Detection. 913 : FasterRCNN Demo Model Inference FRCNN Object Detection. 924 : MobileNet SSD Demo Model Inference MobileNet SSD Object Detection. 935 : TOG Attack Demo Adversarial Attack Real-time TOG Attack against Yolov3 Tiny. 946 : PCB Attack Demo Adversarial Attack Real-time PCB Attack against Yolov3 Tiny. 95``` 96 97Use `what example run` to run examples. 98 99``` 100 Demo Type Description 101-------------------------------------------------------------------------------- 1021 : Yolov3 Demo Model Inference Yolov3 Object Detection. 1032 : Yolov4 Demo Model Inference Yolov4 Object Detection. 1043 : FasterRCNN Demo Model Inference FRCNN Object Detection. 1054 : MobileNet SSD Demo Model Inference MobileNet SSD Object Detection. 1065 : TOG Attack Demo Adversarial Attack Real-time TOG Attack against Yolov3 Tiny. 1076 : PCB Attack Demo Adversarial Attack Real-time PCB Attack against Yolov3 Tiny. 108 109Please input the example index: 110``` 111 112<br /> 113 114## what.utils 115 116This module implements several utility functions. 117 118<br /> 119 120''' 121 122# Project Imports 123from what import models 124from what import attacks 125from what import utils 126 127# Semantic Version 128__version__ = "0.2.2"