Recovering Detail in 3D Shapes Using Disparity Maps

TL;DR

This paper introduces a fine-tuning method that enhances 3D shape details reconstructed from single images by transforming disparity maps into point clouds and integrating them with existing data, improving geometric fidelity.

Contribution

The paper presents a novel approach to convert monocular disparity maps into detailed 3D point clouds using shape priors and optimization, enhancing 3D reconstructions from single images.

Findings

Improved detail in 3D reconstructions demonstrated on synthetic images.

Effective integration of disparity-based point clouds with existing geometries.

Method outperforms baseline approaches in fidelity of 3D shape recovery.

Abstract

We present a fine-tuning method to improve the appearance of 3D geometries reconstructed from single images. We leverage advances in monocular depth estimation to obtain disparity maps and present a novel approach to transforming 2D normalized disparity maps into 3D point clouds by using shape priors to solve an optimization on the relevant camera parameters. After creating a 3D point cloud from disparity, we introduce a method to combine the new point cloud with existing information to form a more faithful and detailed final geometry. We demonstrate the efficacy of our approach with multiple experiments on both synthetic and real images.