Comparison of monocular depth estimation methods using geometrically relevant metrics on the IBims-1 dataset

Tobias Koch, Lukas Liebel, Marco Körner, Friedrich Fraundorfer

Research output: Contribution to journalArticleResearchpeer-review

Abstract

The task of predicting a dense depth map from a monocular RGB image, commonly known as single-image depth estimation (SIDE) or monocular depth estimation (MDE), is an active research topic in computer vision for decades. With the significant progress of deep models in recent years, new standards were set yielding remarkable results in capturing the 3D structure from a single image. However, established evaluation schemes of predicted depth maps are still limited, as they only consider global statistics of the depth residuals. In order to allow for a geometry-aware analysis, we propose a set of novel quality criteria addressing the preservation of depth discontinuities and planar regions, the depth consistency across the image, and a distance-related assessment. As current datasets do not fulfill the requirements of all proposed error metrics, we provide a new high-quality indoor RGB-D test dataset, acquired by a digital single-lens reflex (DSLR) camera together with a laser scanner. New insights into the performance of current state-of-the-art SIDE approaches, as well as subtle differences among them, could be unveiled by employing the proposed error metrics on our reference dataset. Additionally, investigations on the real-world applicability of SIDE methods by a series of experiments regarding different image augmentations, illumination changes and textured planar regions have shown current limitations in this research field.

Original languageEnglish
Article number102877
JournalComputer vision and image understanding
DOIs
Publication statusPublished - 1 Jan 2019

Fingerprint

Camera lenses
Computer vision
Lighting
Statistics
Geometry
Lasers
Experiments

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition

Cite this

Comparison of monocular depth estimation methods using geometrically relevant metrics on the IBims-1 dataset. / Koch, Tobias; Liebel, Lukas; Körner, Marco; Fraundorfer, Friedrich.

In: Computer vision and image understanding, 01.01.2019.

Research output: Contribution to journalArticleResearchpeer-review

@article{a75401a15a5749f081bc6452ffafd843,
title = "Comparison of monocular depth estimation methods using geometrically relevant metrics on the IBims-1 dataset",
abstract = "The task of predicting a dense depth map from a monocular RGB image, commonly known as single-image depth estimation (SIDE) or monocular depth estimation (MDE), is an active research topic in computer vision for decades. With the significant progress of deep models in recent years, new standards were set yielding remarkable results in capturing the 3D structure from a single image. However, established evaluation schemes of predicted depth maps are still limited, as they only consider global statistics of the depth residuals. In order to allow for a geometry-aware analysis, we propose a set of novel quality criteria addressing the preservation of depth discontinuities and planar regions, the depth consistency across the image, and a distance-related assessment. As current datasets do not fulfill the requirements of all proposed error metrics, we provide a new high-quality indoor RGB-D test dataset, acquired by a digital single-lens reflex (DSLR) camera together with a laser scanner. New insights into the performance of current state-of-the-art SIDE approaches, as well as subtle differences among them, could be unveiled by employing the proposed error metrics on our reference dataset. Additionally, investigations on the real-world applicability of SIDE methods by a series of experiments regarding different image augmentations, illumination changes and textured planar regions have shown current limitations in this research field.",
author = "Tobias Koch and Lukas Liebel and Marco K{\"o}rner and Friedrich Fraundorfer",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.cviu.2019.102877",
language = "English",
journal = "Computer vision and image understanding",
issn = "1077-3142",
publisher = "Academic Press",

}

TY - JOUR

T1 - Comparison of monocular depth estimation methods using geometrically relevant metrics on the IBims-1 dataset

AU - Koch, Tobias

AU - Liebel, Lukas

AU - Körner, Marco

AU - Fraundorfer, Friedrich

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The task of predicting a dense depth map from a monocular RGB image, commonly known as single-image depth estimation (SIDE) or monocular depth estimation (MDE), is an active research topic in computer vision for decades. With the significant progress of deep models in recent years, new standards were set yielding remarkable results in capturing the 3D structure from a single image. However, established evaluation schemes of predicted depth maps are still limited, as they only consider global statistics of the depth residuals. In order to allow for a geometry-aware analysis, we propose a set of novel quality criteria addressing the preservation of depth discontinuities and planar regions, the depth consistency across the image, and a distance-related assessment. As current datasets do not fulfill the requirements of all proposed error metrics, we provide a new high-quality indoor RGB-D test dataset, acquired by a digital single-lens reflex (DSLR) camera together with a laser scanner. New insights into the performance of current state-of-the-art SIDE approaches, as well as subtle differences among them, could be unveiled by employing the proposed error metrics on our reference dataset. Additionally, investigations on the real-world applicability of SIDE methods by a series of experiments regarding different image augmentations, illumination changes and textured planar regions have shown current limitations in this research field.

AB - The task of predicting a dense depth map from a monocular RGB image, commonly known as single-image depth estimation (SIDE) or monocular depth estimation (MDE), is an active research topic in computer vision for decades. With the significant progress of deep models in recent years, new standards were set yielding remarkable results in capturing the 3D structure from a single image. However, established evaluation schemes of predicted depth maps are still limited, as they only consider global statistics of the depth residuals. In order to allow for a geometry-aware analysis, we propose a set of novel quality criteria addressing the preservation of depth discontinuities and planar regions, the depth consistency across the image, and a distance-related assessment. As current datasets do not fulfill the requirements of all proposed error metrics, we provide a new high-quality indoor RGB-D test dataset, acquired by a digital single-lens reflex (DSLR) camera together with a laser scanner. New insights into the performance of current state-of-the-art SIDE approaches, as well as subtle differences among them, could be unveiled by employing the proposed error metrics on our reference dataset. Additionally, investigations on the real-world applicability of SIDE methods by a series of experiments regarding different image augmentations, illumination changes and textured planar regions have shown current limitations in this research field.

UR - http://www.scopus.com/inward/record.url?scp=85076557147&partnerID=8YFLogxK

U2 - 10.1016/j.cviu.2019.102877

DO - 10.1016/j.cviu.2019.102877

M3 - Article

JO - Computer vision and image understanding

JF - Computer vision and image understanding

SN - 1077-3142

M1 - 102877

ER -