notesum.ai

Published at November 7

D$^3$epth: Self-Supervised Depth Estimation with Dynamic Mask in Dynamic Scenes

cs.CV

cs.AI

cs.LG

Released Date: November 7, 2024

Authors: Siyu Chen, Hong Liu, Wenhao Li, Ying Zhu, Guoquan Wang, Jianbing Wu

Arxiv: http://arxiv.org/abs/2411.04826v1

Refer to caption

Method	Test Frames	Semantic	W $\times$ H	Errors $\downarrow$				Accuracy $\uparrow$
Method	Test Frames	Semantic	W $\times$ H	AbsRel	SqRel	RMSE	RMSE_log	$\delta<1.25$	$\delta<1.25^{2}$	$\delta<1.25^{3}$
Struct2Depth [23]	1	$\bullet$	416 $\times$ 128	0.141	1.026	5.291	0.215	0.816	0.945	0.979
Bian et al. [24]	1		416 $\times$ 128	0.137	1.089	5.439	0.217	0.830	0.942	0.975
Gordon et al. [25]	1	$\bullet$	416 $\times$ 128	0.128	0.959	5.230	0.212	0.845	0.947	0.976
Monodepth2 [11]	1		640 $\times$ 192	0.115	0.903	4.863	0.193	0.877	0.959	0.981
ManyDepth [8]	2(-1,0)		640 $\times$ 192	0.098	0.770	4.459	0.176	0.900	0.965	0.983
DynamicDepth [9]	2(-1,0)	$\bullet$	640 $\times$ 192	0.096	0.720	4.458	0.175	0.897	0.964	0.984
DepthFormer [20]	2(-1,0)		640 $\times$ 192	0.090	0.661	4.149	0.175	0.905	0.967	0.984
DualRefine [10]	2(-1,0)		640 $\times$ 192	0.090	0.658	4.237	0.171	0.912	0.967	0.984
MOVEDepth [26]	2(-1,0)		640 $\times$ 192	0.089	0.663	4.216	0.169	0.904	0.966	0.984
MonoViFI [27]	3(-1,0,1)		640 $\times$ 192	0.091	0.589	4.088	0.166	0.912	0.969	0.986
Baseline [28]	2(-1,0)		640 $\times$ 192	0.086	0.613	4.096	0.165	0.915	0.969	0.985
Ours	2(-1,0)		640 $\times$ 192	0.084	0.607	4.086	0.164	0.917	0.969	0.985