AI RESEARCH
Block-Sparse Global Attention for Efficient Multi-View Geometry Transformers
arXiv CS.CV
•
ArXi:2509.07120v2 Announce Type: replace Efficient and accurate feed-forward multi-view reconstruction has long been an important task in computer vision. Recent transformer-based models like VGGT, $\pi^3$ and MapAnything have nstrated remarkable performance with relatively simple architectures. However, their scalability is fundamentally constrained by the quadratic complexity of global attention, which imposes a significant runtime bottleneck when processing large image sets.