Skip to main content

Movement Detector

Detect abrupt movement in videos

Project-Scoped Page

The Movement Detector is designed to detect abrupt movement in videos. This is used to quickly detect the onset of behaviors, which are then manually confirmed in Quick Score.

Algorithms

The Movement Detector currently supports a single algorithm: Intensity Change.

If you are in need of a different algorithm, please let us know at support@codeneuro.com!

Parameters

To learn more about the currently supported parameters (threshold, small object removal, opening, and closing), go to the Intensity Change algorithm page.

Time-based ROI

You have the option to create one or more region-of-interests (ROIs) to constrain the algorithm to.

This is useful for two reasons:

  1. A smaller ROI means a faster analysis

  2. An ROI will constrain the ‘noise’ from pixels unrelated to the subject or bodypart of interest.

Create an ROI

To create an ROI, simply click and drag a rectangle on the video.

First ROI example

In the timeline below the video, you will see an indicator for the first ROI appear. Notice that the first ROI is always anchored to the beginning of the video.

As you progress through the video, you can click the + button on the timeline to create a new ROI at that timepoint.

Second ROI example

Shift an ROI

You can shift the onset time of an ROI by dragging the indicator on the timeline. (You cannot shift the time of the first ROI.)

Shift an ROI

Delete an ROI

To delete an ROI, hover over the ROI indicator on the timeline and click the trash icon.

Delete an ROI

Bypass Ranges

For longer videos, you may want to skip over certain timepoints.

For example, when studying sleep, you may want to skip over the wake periods, which will generate far larger movements than the sleep periods.

To bypass time ranges in the video, simply upload a CSV file in the following format:

start_time (in seconds)end_time (in seconds)
010
1020

Requirements

The CSV must have two columns—a start and end time (in seconds). The labels of the columns do not matter.

A start time must be less than its corresponding end time.

Each row must be in ascending order, and cannot overlap.

Example of an invalid CSV:

start_time (in seconds)end_time (in seconds)
010
5

(wrong!)

20

Video Previews

Bypass ranges will be ignored in video previews.

Analysis Preview

To preview the analysis before running it, you can click the Preview button in the settings panel.

Video preview button

To start the preview from the current timepoint of the video above, click Set Start Time.

This will generate a 5-second preview of the analyzed video and accompanying timeseries.

analysis preview