Towards Consistent Video Geometry Estimation 事件

Towards Consistent Video Geometry Estimation arXiv:2605.30060v1 Announce Type: new Abstract: This work presents ViGeo, a feed-forward foundation model for recovering spatially dense and temporally consistent geometry from video sequences. Built upon a plain transformer architecture without task-specific architectural modifications, ViGeo supports streaming, full-sequence, and long-video inference within a unified model. The key design is dynamic chunking attention, which exposes the model to bo