A Multi-view Stereo Benchmark with High-Resolution Images and Multi-camera Videos

TLDR

Current multi‑view stereo benchmarks lack diverse scenes and high‑resolution imagery, limiting their usefulness for realistic reconstruction. This work introduces a new multi‑view stereo dataset designed to address those limitations. The dataset was built by capturing indoor and outdoor scenes with a high‑precision laser scanner, high‑resolution DSLR images, and synchronized low‑resolution stereo videos, and aligning them to the scans using a photometric‑error minimization technique conditioned on geometry. The resulting benchmark offers diverse viewpoints, high temporal and spatial resolution, supports hand‑held mobile devices, and is publicly available with an online evaluation server.

Abstract

Motivated by the limitations of existing multi-view stereo benchmarks, we present a novel dataset for this task. Towards this goal, we recorded a variety of indoor and outdoor scenes using a high-precision laser scanner and captured both high-resolution DSLR imagery as well as synchronized low-resolution stereo videos with varying fields-of-view. To align the images with the laser scans, we propose a robust technique which minimizes photometric errors conditioned on the geometry. In contrast to previous datasets, our benchmark provides novel challenges and covers a diverse set of viewpoints and scene types, ranging from natural scenes to man-made indoor and outdoor environments. Furthermore, we provide data at significantly higher temporal and spatial resolution. Our benchmark is the first to cover the important use case of hand-held mobile devices while also providing high-resolution DSLR camera images. We make our datasets and an online evaluation server available at http://www.eth3d.net.