High Fidelity 3D Reconstructions with Limited Physical Views

TL;DR

This paper introduces a neural shape prior-based method that achieves high-fidelity 3D reconstructions from only 2-3 uncalibrated views, reducing the need for costly multi-camera setups.

Contribution

It presents a novel approach combining 2D-3D lifting with multi-view equivariance to enable high-quality 3D reconstructions from limited views.

Findings

Achieves comparable fidelity to multi-camera rigs with only 2-3 uncalibrated views.

Leverages neural shape priors and multi-view equivariance for improved reconstruction.

Reduces hardware costs and complexity in 3D reconstruction setups.

Abstract

Multi-view triangulation is the gold standard for 3D reconstruction from 2D correspondences given known calibration and sufficient views. However in practice, expensive multi-view setups -- involving tens sometimes hundreds of cameras -- are required in order to obtain the high fidelity 3D reconstructions necessary for many modern applications. In this paper we present a novel approach that leverages recent advances in 2D-3D lifting using neural shape priors while also enforcing multi-view equivariance. We show how our method can achieve comparable fidelity to expensive calibrated multi-view rigs using a limited (2-3) number of uncalibrated camera views.