Semantic data augmentation

• Based on numpy, OpenCV, picking the best from each of them.
• Simple, flexible API that allows the library to be used in any computer vision pipeline.
• Easy to extend the library to wrap around other libraries.
• Easy to extend to other tasks.
• Supports python 2.7-3.7
• Easy integration with PyTorch.
• Supports extraction of people on segmented images.

Table of contents

• How to use
• Installation
  • People extract
  • Overlay road with people

How to use

• clone the repository

git clone https://github.com/nemodrive/semantic-data-augmentation.git

• download Cityscapes Dataset from Cityscape Dataset.
• create a two columns CSV file with original image path and coresponding segmented image path (one example is in resources/good_train_fine.txt)
• create your own dataset (similar to Cityscape Dataset), having the original image and segmented road of that
• have fun :)

Installation

1. Install pip

curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py
python get-pip.py

Extract people

2. Clone the repository

git clone https://github.com/nemodrive/semantic-data-augmentation.git

3. Go to extract_people.py

cd scripts

4. Run extract_people.py

python extract_people.py <path_to_CSV_file>

Overlay road with people

5. You can install roadpackage using pip command

pip install git+https://github.com/nemodrive/semantic-data-augmentation.git

6. Import in your file

from roadpackage.road import overlay_people_on_road