Addressing the pain point that "text only provides high-level semantics, while keyframes provide precise spatial-temporal constraints but are difficult to coordinate with text," this paper proposes a ...

• Linear: Changes between keyframes at a constant rate—simple and linear movement. It moves at a constant speed from start to finish, with an abrupt start/stop at the points (Abrupt start/stop, ...

description [CVPR 2025][Video Understanding][Long Video Understanding] This paper proposes DrVideo, which reformulates **long video understanding as a long document understanding** task: it first ...