Booster: a Benchmark for Depth from Images of Specular and Transparent Surfaces

TL;DR

This paper introduces Booster, a high-resolution dataset with dense ground-truth labels for depth estimation on specular and transparent surfaces, addressing challenges in non-Lambertian material handling and high-res image processing.

Contribution

It presents a novel dataset with accurate labels, including high-res and unbalanced stereo pairs, and a new acquisition pipeline using deep space-time stereo for precise labeling.

Findings

Dataset reveals open challenges in depth estimation for non-Lambertian surfaces.

Experiments demonstrate current limitations and future research directions.

Provides benchmarks for stereo and monocular depth estimation methods.

Abstract

Estimating depth from images nowadays yields outstanding results, both in terms of in-domain accuracy and generalization. However, we identify two main challenges that remain open in this field: dealing with non-Lambertian materials and effectively processing high-resolution images. Purposely, we propose a novel dataset that includes accurate and dense ground-truth labels at high resolution, featuring scenes containing several specular and transparent surfaces. Our acquisition pipeline leverages a novel deep space-time stereo framework, enabling easy and accurate labeling with sub-pixel precision. The dataset is composed of 606 samples collected in 85 different scenes, each sample includes both a high-resolution pair (12 Mpx) as well as an unbalanced stereo pair (Left: 12 Mpx, Right: 1.1 Mpx), typical of modern mobile devices that mount sensors with different resolutions. Additionally, we provide manually annotated material segmentation masks and 15K unlabeled samples. The dataset is composed of a train set and two test sets, the latter devoted to the evaluation of stereo and monocular depth estimation networks. Our experiments highlight the open challenges and future research directions in this field.