Video Generator

Visualization only

The video generator only draws the tracking results. To correct them, use the Validator.

The video generator renders videos of your tracking results, to check them visually or to share them. Launch it with:

idtrackerai_video

or open a session directly with idtrackerai_video path/to/session_folder.

idtracker.ai’s video generator application (in dark mode)

idtracker.ai’s video generator application (in light mode)

It can generate two types of videos, shown in the examples below (left to right):

Pose estimation

Individual videos are a good input for pose estimation tools such as SLEAP and DeepLabCut.

• General video: the original video with the trajectories and, optionally, the identity labels drawn on top (left).

• Individual videos: one video per animal, cropped around it and following it over time (center), plus a collage of all individual videos in a grid (right).

Tip

Use the resize option of the general video to make the output smaller. Smaller videos are faster to generate and easier to store and share.

Command-line interface

To generate a video without the graphical interface, add --run:

idtrackerai_video path/to/session_folder --run

Option	Description
--run	Generate the video without the GUI.
--individual	Generate individual videos instead of a general video.
--gray	Draw the original video in grayscale.
--t <path>	Trajectory file to draw. Default: session_folder/trajectories/trajectories*.
--tl <int>	Trail length: number of past positions drawn for each animal in general videos. Default: 20.
--s <int>	First frame of the video.
--e <int>	Last frame of the video.
--size <int>	Side of the square individual videos, in pixels. Default: the median body length of the animals.
--no-labels	Draw centroids without identity labels in general videos.

Run idtrackerai_video -h to print this list. To generate videos from Python, see idtrackerai.extra_tools.video_generator.