Benchmark Radar
AI BENCHMARK PROFILE

CamChoreo

General AIMultimodal PerceptionCamChoreo Benchmark Team

Evaluates temporally grounded compositional camera motion understanding using 4,229 single-shot clips with expert-annotated segments and 20 direction-aware labels.

Released
2026-08-11
Readiness
Paper only
Primary field
General AI

Why it matters

Advances camera motion evaluation beyond clip-level labels to compositional and temporal localization, exposing limitations in current multimodal models.

Motivation

Understanding camera motion is fundamental to video perception, with applications in spatial intelligence and controllable video generation.

Primary resources

Benchmark Radar records only publicly supported details and links back to primary sources for verification.