TABLE 1.

Quantitative study on proposed depth model on UCL 2019 synthetic data. Best values are in bold, and the second best values are underlined. Abs Rel, Sq Rel, RMSE, and RMSE log are in mm, and δ < 1.25¹ , δ < 1.25² , and δ < 1.25³ are in percentage.

Model	Abs Rel ↓	Sq Rel ↓	RMSE ↓	RMSE log ↓	δ < 1.251 ↑	δ < 1.252 ↑	δ < 1.253 ↑
SfMLearner (Zhou et al., 2017)	0.451	0.711	1.768	0.537	0.366	0.618	0.785
Monodepth1 (Godard et al., 2017)	0.444	0.752	1.755	0.531	0.371	0.625	0.790
DDVO (Wang et al., 2018)	0.452	0.772	1.768	0.538	0.366	0.618	0.785
Monodepth2 (Godard et al., 2019)	0.446	0.756	1.757	0.532	0.360	0.624	0.789
HR Depth (Lyu et al., 2021)	0.448	0.762	1.762	0.534	0.369	0.621	0.788
AF-SfMLearner (Shao et al., 2021)	0.387	0.618	1.648	0.480	0.420	0.686	0.831
AF-SfMLearner2 (Shao et al., 2022)	0.352	0.545	1.581	0.448	0.445	0.712	0.852
Our	0.141	0.115	0.669	0.179	0.828	0.959	0.985