Depth Estimation from a Single Image Using Line Segments only

被引：0

作者：

Nava Zavala, Jose G. ^{[1
]}

Martinez-Carranza, Jose ^{[1
]}

机构：

[1] Inst Nacl Astrofis Opt & Electr, Luis Enrique Erro 1, Puebla, Mexico

来源：

ADVANCES IN ARTIFICIAL INTELLIGENCE-IBERAMIA 2022 | 2022年 / 13788卷

关键词：

Depth estimation; Monocular camera; Line segments;

D O I：

10.1007/978-3-031-22419-5_28

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We present a method for depth estimation from a single image using an intermediate representation in the form of line segments. Rather than regressing depth from a chromatic image in RGB format, we explore the use an image containing line segments extracted from the original chromatic image using the Line Segment Detector (LSD), arguing that this image, even when sparse in visual data, still contains information to infer a depth image. Our proposed approach has been tested on the NYU-depth dataset for indoor scenes and on simulated images created with Airsim, seeking to assess the performance of our method with synthetic images. Our experiments show promising results confirming that it is possible to estimate a depth image from a single image containing line segments only.

引用

页码：331 / 341

页数：11

共 24 条

[1]

Alhashim I, 2019, Arxiv, DOI arXiv:1812.11941

[2] Estimating Depth From Monocular Images as Classification Using Deep Fully Convolutional Residual Networks [J].

Cao, Yuanzhouhan ;

Wu, Zifeng ;

Shen, Chunhua .

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2018, 28 (11) :3174-3182

[3]

Chen SY, 2020, Arxiv, DOI arXiv:2006.01047

[4]

Eigen D, 2014, ADV NEUR IN, V27

[5]

Fu H, 2018, Arxiv, DOI arXiv:1806.02446

[6]

Garcia-Garcia A, 2017, Arxiv, DOI [arXiv:1704.06857, DOI 10.48550/ARXIV.1704.0685729, 10.48550/arXiv.1704.06857]

[7]

Geiger A, 2012, PROC CVPR IEEE, P3354, DOI 10.1109/CVPR.2012.6248074

[8]

Hao ZX, 2018, Arxiv, DOI arXiv:1809.00646

[9]

Jiang H, 2021, ARXIV

[10] DepthNet: A Recurrent Neural Network Architecture for Monocular Depth Prediction [J].

Kumar, Arun C. S. ;

Bhandarkar, Suchendra M. ;

Prasad, Mukta .

PROCEEDINGS 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW), 2018, :396-404

← 1 2 3 →