Skip to content

usc-auv-team/cv.annotation

Repository files navigation

Image and Video Annotation Tool

@author: frozenscrypt

Tool is under maintenance and feature additions with video annotation script coming soon!

Uses python 3

Installations:

For ubuntu users

  • sudo apt-get install python-tk
  • sudo apt-get isntall opencv-python
  • pip install Pillow

For windows users, exe file will be updated shortly!

Instructions:

  • GUI utilizes images numbered starting from 1

  • The bash script renames all images in a directory as needed. run rename_images.sh

  • To run the GUI python image_annotation_gui.py path-to-configFile

Config-file is a json file with keys to specify the filename(input folder directory) , image folder name, output folder name, multiplier to scale the images as needed(<1), frame number or the initial image number( tells the gui from where to start the annotation process)

Features:

  • Allows to draw multiple boxes to a image
  • Each time a box is drawn and a comment is added if needed, click submit to save the annotation
  • Clear button allows to clear a faulty annotation but once submitted, it can't be cleared
  • Next button calls the next image
  • Previous button calls the previous image and deletes the previous annotation from the data file
  • Log box displays the current operation on the current image

About

Annotation Tool for Tensorflow Training

Resources

Stars

Watchers

Forks

Packages

No packages published