One of the most common questions we get is "How can I use computer vision object detection models with video?" The answer is simple: you treat each frame as an image.
But, in practice, that has been quite cumbersome. It's meant fumbling around with command line tools, experimenting with various options, and manually selecting frames.
We wanted to make it easier, so we added the ability to drop video files directly into Roboflow Pro. You simply select a sample rate (since you'll want to label just a subset of the thousands of frames in your video). Then it's off to the races; use Roboflow Pro's integrated labeling integrations to annotate the video frames yourself or outsource the job with just a couple of clicks.
We're excited to see what you build! We've already seen use-cases in sports, autonomous vehicles, and CCTV.