Video Annotation

Video Annotation Intro Video

Learn how easy it is to maintain context between objects.

Introduction

The motivations for using video include:

  • Time Series information. I.e. a Car changing lanes that happens over multiple frames.
  • Increase volume of annotations.
  • Reduce search space on what needs to be labelled. I.e. Scan through a video to pick out a rare event.

📘

Latest version of Chrome browser required

More annotation client specs: Minimum Specs.

Learning Overview

Annotating video is a complex and challenging concept.
Like video editing (ie Adobe Premier Pro) there are many concepts to learn to get effective results.
Unlike video editing, Deep Learning is new ground , so there are new concepts about to learn in addition to the specific interface implementation.

We suggest watching the above video and reading (or at least skimming).

📘

Brace yourself - there's a lot to learn! You can do it!

Suggested reading

Annotators:

  • Video Annotation Concepts
  • What a Sequence is
  • Interpolation (Optional)

Technical and Admins:

  • All of Annotators items
  • Video Import Specs

After reading play around in a test project to get a feel for it.