Computer Vision Annotation Tool, or CVAT for short, is an Open Source tool used for labeling image and video data for computer vision algorithms. It is one of the most popular Open Source annotation tools out there and makes video annotations easier and faster. This workshop will use the online demo version of CVAT so that we don’t have to worry about any local software installations. We will leverage this plug and play tool to bring in raw data, set up an annotation project, understand the how’s and why’s of choosing to do data annotations a particular way, establish best practices that we will follow for the workshop, leverage ways to speed up the annotation process, export the resulting annotations, and understand ways to gain meaningful insights into how to evaluate current data annotations. We will also highlight how CVAT may be leveraged in a production environment and how to take the data annotation process beyond a one-person effort into a team-coordinated and validated pipeline where one can continually iterate and improve their annotations. Dare to strike gold and derive meaning out of your own raw data in this workshop!
- Intro - 5 minutes
- CVAT project setup - 10 minutes
- Read annotations specifications - 5 minutes
- Read annotations best practices - 5 minutes
- Follow the annotation steps to annotate video - 30 minutes
- Conclusion - 5 minutes
- Q & A - throughout
- Laptop & charger
- Internet access
- Understand the challenges and basics of performing successful data annotations for visual data.
- Gain useful insights into how a professional data annotation team performs, questions they ask as they work through the data annotation process, and best practices they abide by.
- Engage and connect through team working skills and workshop interactions while exploring new areas of expertise with new people.