Visualization

This modules provides an interface to visualise and save video clips or frames of sign language datasets. The video could be enriched by annotations and overlays such as pose information. The annotations linked to the video could also be displayed liked showed in this example :

Example of video player

Hands on example of how to use the video player could be found in the examples folder in the project repository. Once the player is launched, you can:

  • Use S key to save the current frame in the screenshot folder
  • Use Space to play and pause the current video
  • Use the Arrows to move back and forth in the video
  • Use Q to exit the video player

The VideoPlayer allows you to configure the result you want to see by attaching information. The object that could be attached are the following :

  • Video: Only one video could be attached to the player. The video could either be a video file or a folder containing one image for each frame of the video.
  • Pose: You can attach as many pose as you like to a player. If you are using Mediapipe of OpenPose, the pose module provides some useful functions to pre-process or modify those landmarks.
  • Segments: A segment is represented by a start timestamp, an end timestamp and a label. Segments represents annotations associated with the video and will be displayed along with it.

For more information about the VideoPlayer object and its methods refer to the reference