Over the years, we’ve seen thousands of companies doing video labeling. But as diverse as these use cases are on the surface, we've found that they all fit somewhere on a spectrum of labeling specificity, from broad to granular:
At the broad end of the spectrum, we have video classification use cases. This is where annotators classify an entire video. Is it NSFW? Is it clear? What is the topic of the video?
At the most detailed end of the spectrum, we have video object detection and video object tracking, which involves identifying and tracking objects in every frame of a video. This is particularly helpful for cases where you want to track moving objects for computer vision applications. For example, tracking a car, a shopper, or an athlete.
Today, we’re excited to announce a new feature that addresses the center of the spectrum: video frame classification.
Check out this video from ML Evangelist Micaela Kaplan to see video frame classification in action!
This new feature makes it easy to classify individual frames of a video. For example, is there motion, slow motion, or no movement? Is the robot arm grabbing - or releasing?
If you’re interested in implementing this new capability, check out our docs to get started. This is available for all our Label Studio Enterprise Cloud customers today.
Interested in using Label Studio Enterprise for your image labeling use case? We’d love to chat! You can schedule some time to talk here.
We’ve released a new version of the YOLO ML backend connector designed to support YOLOv8, which now supports advanced object detection, segmentation, classification, and video object tracking with Label Studio.
We’ve updated the reviewer workflow to make it easier and more intuitive.
Label Studio is already the most customizable labeling platform. We’re making it even more flexible with custom scripts.
