Indoor Monocular Depth Estimation Based on Global-Local Feature Fusion

Fig. 9. Comparison of depth maps predicted by each method in different indoor scenes. (a) LapDepth; (b) GLPDepth; (c) proposed method; (d) ground truth

Download full size

Fig. 10. Comparison of 3D reconstruction results for each method in different indoor scenes. (a) LapDepth; (b) GLPDepth; (c) proposed method; (d) ground truth

Download full size

Fig. 11. Ablation visualization results of the MKCM

Download full size

Table 1. Quantitative comparison of different methods on the NYU Depth V2 dataset

View table

Table 1. Quantitative comparison of different methods on the NYU Depth V2 dataset

Method	Params /10⁶	Abs Rel	RMSE	$l o g R M S E$	$l o g M A E$	$δ < 1.25$	$δ < 1 . 25^{2}$	$δ < 1 . 25^{3}$
Method of reference ［35］	141	0.158	0.641	0.214		0.769	0.950	0.988
BTS^［21］	47	0.110	0.392	0.142	0.047	0.885	0.978	0.994
AdaBins^［36］	78	0.103	0.364		0.044	0.903	0.984	0.997
LapDepth^［25］	73	0.108	0.385	0.137	0.046	0.890	0.982	0.996
GLPDepth^［27］	62	0.106	0.375	0.133	0.045	0.895	0.984	0.996
Method of reference ［37］		0.107	0.389		0.042	0.892	0.987	0.998
Proposed method	61	0.100	0.361	0.127	0.043	0.907	0.986	0.997

Table 2. The ablation experimental results of the proposed method on the NYU Depth V2 dataset

View table

Table 2. The ablation experimental results of the proposed method on the NYU Depth V2 dataset

A	B	C	Runing time /ms	Abs Rel	RMSE	$l o g R M S E$	$l o g M A E$	$δ < 1.25$	$δ < 1 . 25^{2}$	$δ < 1 . 25^{3}$
×	×	×	18.2	0.105	0.376	0.133	0.045	0.897	0.984	0.996
√	×	×	18.5	0.102	0.369	0.130	0.044	0.903	0.986	0.997
×	√	×	22.7	0.103	0.362	0.129	0.044	0.901	0.985	0.997
×	×	√	18.7	0.102	0.370	0.131	0.045	0.905	0.985	0.997
√	√	×	22.9	0.100	0.359	0.127	0.043	0.906	0.987	0.997
√	×	√	18.7	0.101	0.365	0.129	0.044	0.908	0.987	0.997
×	√	√	22.9	0.102	0.366	0.129	0.044	0.903	0.986	0.997
√	√	√	23.1	0.100	0.361	0.127	0.043	0.907	0.986	0.997

Tools

Get Citation

Copy Citation Text

Zengyu Tian, Changku Sun, Yue Li, Luhua Fu, Peng Wang. Indoor Monocular Depth Estimation Based on Global-Local Feature Fusion[J]. Laser & Optoelectronics Progress, 2025, 62(14): 1412004

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Instrumentation, Measurement and Metrology

Received: Jan. 2, 2025

Accepted: Feb. 25, 2025

Published Online: Jul. 16, 2025

The Author Email: Peng Wang (wang_peng@tju.edu.cn)

DOI:10.3788/LOP250436

CSTR:32186.14.LOP250436

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology