Skip to main content

End to end software to capture new objects using RGBD camera and augment them to get a annotated dataset to train deep nets.

Project description

Easy Augmentor

End to end software to capture new objects using RGBD camera and augment them to get a annotated dataset to train deep nets

  1. Pipeline to artificial generate annotated data for training deep learning models.
  2. Pipeline includes starting from capturing images using provided camera (Realsense), generate semantic labels of the captured image and then generate the artificial images.
  3. GUI required for the end user from capturing to labelling and generating new data.

Requirements

  1. Ubuntu 16.04 (Testing for Ubuntu 18.04)
  2. Intel Realsense Camera
  3. Processer intel i5 or higher
  4. libpcl-dev==1.7
  5. python 3.5

Limitations

  1. Number of classes captured should be more than or equal to 2.

Installation

Linux:

pip3 install easy-augmentor

Usage example

cd realsense_augmentor
python3 src/main.py
  1. Start page will be as below and select the path to save the captured images. alt text
  2. On selecting capture next window will look like alt text
  3. On selectin Have Annotations alt text Steps after selecting capture option and selecting save path
  4. Next window will have image and mask of corresponding object. Capture as many as images per classe. alt text
  5. Click add to add new class label. alt text
  6. Click save to save the current displyed image and semantic label.
  7. Click finish once done with capturing all the images.
  8. Folder name \textbf{captured_data} in selected save path will have images, labels and labels.txt
  9. Input parameters to generate artificial images need to be filled next window. alt text
  10. Click OK once setting up the parameters.
  11. Folder name \textbf{augmented} in selected save path will have artificial images.

Steps after selecting Have annotations 12) Select the location of images and corresponding annotations 13) Select save path 14) Select labels.txt file path 15) Follow the steps from Step 9

Release History

  • 1.0.0
    • First release for crowd testing

Contributors

Distributed under the MLP license. See LICENSE for more information.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for easy-augment, version 1.0.2
Filename, size File type Python version Upload date Hashes
Filename, size easy_augment-1.0.2-py3-none-any.whl (40.9 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size easy-augment-1.0.2.tar.gz (28.2 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page