Think Locally, Refine Globally for Memory-Efficient 3D Reconstruction

2026-09-18

AI Summary

提出LoG-VGGT框架，通过局部跨窗口注意力和全局相机一致性优化，解决长序列3D重建中的内存效率与姿态漂移问题。

Abstract

We propose LoG-VGGT, a memory-efficient framework for long-sequence 3D reconstruction that balances local temporal modeling with global camera consistency. Instead of relying on full global attention, our method introduces cross-window attention at a small subset of transformer blocks, enabling effective information propagation across adjacent temporal windows while keeping memory usage bounded. To mitigate long-term pose drift, we further design a global camera consistency refinement module, where camera tokens interact with compact register tokens via cross-attention to enforce scene-level constraints across the entire sequence. This design enables joint optimization of camera representations and significantly improves long-horizon pose stability without incurring the high cost of sequence-wide attention. Extensive experiments demonstrate that LoG-VGGT achieves improved depth accuracy and robust camera pose estimation across multiple long-sequence benchmarks, while delivering competitive streaming reconstruction performance.

Problem

• 3D Reconstruction
• Memory-Efficient
• Long-Sequence
• Global Camera Consistency

Innovation

• cross-window attention
• global camera consistency refinement
• memory-efficient 3D reconstruction