Multi-view Pyramid Transformer: Look Coarser to See Broader

TL;DR

The paper introduces Multi-view Pyramid Transformer (MVP), a scalable architecture that efficiently reconstructs large 3D scenes from many images by hierarchically integrating local details and global context.

Contribution

MVP presents a novel dual hierarchy design that enhances efficiency and detail in 3D scene reconstruction from multiple views.

Findings

Achieves state-of-the-art reconstruction quality with 3D Gaussian Splatting.

Demonstrates high efficiency and scalability across diverse datasets.

Effectively balances detail and global context in large scene reconstruction.

Abstract

We propose Multi-view Pyramid Transformer (MVP), a scalable multi-view transformer architecture that directly reconstructs large 3D scenes from tens to hundreds of images in a single forward pass. Drawing on the idea of ``looking broader to see the whole, looking finer to see the details," MVP is built on two core design principles: 1) a local-to-global inter-view hierarchy that gradually broadens the model's perspective from local views to groups and ultimately the full scene, and 2) a fine-to-coarse intra-view hierarchy that starts from detailed spatial representations and progressively aggregates them into compact, information-dense tokens. This dual hierarchy achieves both computational efficiency and representational richness, enabling fast reconstruction of large and complex scenes. We validate MVP on diverse datasets and show that, when coupled with 3D Gaussian Splatting as the underlying 3D representation, it achieves state-of-the-art generalizable reconstruction quality while maintaining high efficiency and scalability across a wide range of view configurations.