PixelMe is a tool which allows users to classify images and videos into different categories. This is to label the dataset for image classification algorithm development.

To classify all the images within a folder simply use the command

python3 PixelMe.py --main_class <YOUR-MAIN-CLASS> --sub_class <SUB-CLASS-1> <SUB-CLASS-2>

To classify frames in the video use

python3 PixelMe.py --main_class <YOUR-MAIN-CLASS> --sub_class <SUB-CLASS-1> <SUB-CLASS-2> --vid

Here I give the example of classifying cars based on brands

python3 PixelMe.py --main_class cars --sub_class toyota lexus honda --vid

• Click on the load button and select the video file
• When the video is playing use the number keys on your key board to label the frame being displayed
  • Press number key 1 for labelling the frame as toyota
  • Press number key 2 for labelling the frame as lexus
  • Press number key 3 for labelling the frame as honda
• The images can then be found in the folder ./data/cars
• Space bar will pause/play the video. (Only works for videos i.e --vid argument is provided)
• Using the right and left arrow key will
  • Go to next/previous image
  • Forward/rewind the video by 10 seconds
• Use the slider to forward and rewind the video, and the buttons at the bottom to control the video playback
• The slider on the right in the video mode allows you to control the playback speed

Head over to the issues tab and feel free to create one

The parts of the video player is based on the Tkinter VLC player