Video Annotation: Labelling Objects Frame by Frame and Across Time
A video is a long run of images. Labelling it means marking the objects in each frame, and often following the same object across frames so the model learns that it is the same thing.
We label video remotely from Chittorgarh, India, for AI teams in India and abroad. We start with a small pilot batch.
Our main page for this service: AI Data Annotation.
What the work involves
Typical tasks:
- 2D and 3D bounding boxes on frames
- Polygon outlines where the shape matters
- Following each object with a consistent ID from frame to frame
- Marking when an object enters, leaves or is hidden
Keeping the cost under control
Labelling every frame is slow and often unnecessary. You can label every few frames, or mark an object at the start and end of a movement and let the tool fill in the frames between. Many tools support this. We use it where the tool allows and check the result by eye.
Which approach fits depends on how fast things move, so we test it on the pilot batch.
How a pilot works
You send a short clip. We agree a written labelling guide, label the clip, and you check it. We agree how accuracy will be measured and the target before we talk about volume.
Frequently asked questions
Do you track the same object across frames?
Yes. Consistent IDs from frame to frame are part of the work. We agree the rules in the guide.
What video formats can you take?
Tell us the format and the tool. We say plainly whether we can work with them after seeing a sample.
How is our data kept private?
We agree who may see it and sign an NDA before the first file is shared.
How much does it cost?
It depends on the length of video, how many objects there are and how often we label. We quote after seeing a sample.
Send us a short clip
A short clip and a description of the labels you need is enough to start. Call +91-86194 70448 or email contact@gtsltd.in. We work remotely, in English, with businesses in India and abroad.