Pix3D: Dataset and Methods for Single-Image 3D Shape Modeling

TL;DR

This paper introduces Pix3D, a large-scale dataset for single-image 3D shape modeling, along with evaluation benchmarks and a novel multi-task model for reconstruction and pose estimation.

Contribution

It provides a comprehensive dataset with pixel-level alignment, standardized evaluation criteria, and a new multi-task learning model for 3D reconstruction and pose estimation.

Findings

Pix3D dataset enables diverse shape-related tasks.

Benchmarking reveals current reconstruction algorithms' performance.

Proposed model achieves state-of-the-art results in reconstruction and pose estimation.

Abstract

We study 3D shape modeling from a single image and make contributions to it in three aspects. First, we present Pix3D, a large-scale benchmark of diverse image-shape pairs with pixel-level 2D-3D alignment. Pix3D has wide applications in shape-related tasks including reconstruction, retrieval, viewpoint estimation, etc. Building such a large-scale dataset, however, is highly challenging; existing datasets either contain only synthetic data, or lack precise alignment between 2D images and 3D shapes, or only have a small number of images. Second, we calibrate the evaluation criteria for 3D shape reconstruction through behavioral studies, and use them to objectively and systematically benchmark cutting-edge reconstruction algorithms on Pix3D. Third, we design a novel model that simultaneously performs 3D reconstruction and pose estimation; our multi-task learning approach achieves state-of-the-art performance on both tasks.