CameraBench: understanding camera motions in any video
Source: https://arxiv.org/abs/2504.15376
Structured fields
{
"checked_at": "2026-10-17T00:00:00.000Z",
"usefulness": "About 3,000 internet videos annotated by experts with a camera-motion primitive taxonomy designed with cinematographers; a vocabulary for tracking, dolly, pan and so on, and evidence on annotator training.",
"source_type": "paper",
"availability": "available",
"source_owner": "Lin, Mitra et al.",
"next_check_at": "2026-10-24T00:00:00.000Z",
"known_limitations": [
"Official code repository was not found in this pass."
]
}Evidence claims
- [sourced_fact] Abstract: CameraBench has ~3,000 videos annotated by experts through multi-stage quality control, with a taxonomy of camera motion primitives developed with cinematographers. · locator: abstract source
- [upstream_claim] Abstract: finds some motions such as 'follow' require understanding scene content, and that expertise and tutorial training improve annotation accuracy. · locator: abstract source
- [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| documentation | unknown | unknown | unknown | unknown | unknown | arXiv non-exclusive distribution license. Dataset/code release terms not checked. |