EPCDepth

EPCDepth

Results on KITTI raw test set

Backbone	Resolution	Sup	PP	Train	Abs Rel.	Sq Rel.	RMSE	RMSElog	A1	Model
Res50	320x1024	Stereo	gpp	Reported	0.091	0.646	4.207	0.176	0.901	-
Res50	320x1024	Stereo	gpp	Official	0.095	0.667	4.239	0.183	0.889	Baidu/Google
Res50	320x1024	Stereo		Trained	0.090	0.682	4.282	0.178	0.903	Baidu/Google
Res50	320x1024	Stereo	gpp	Trained	0.089	0.660	4.224	0.176	0.905	Baidu/Google

Results on NYU v2 test set (trained on KITTI)

Backbone	Resolution	Sup	PP	Train	Abs Rel.	Sq Rel.	RMSE	RMSElog	A1	Model
Res50	288x384	Stereo	gpp	Reported	0.247	0.277	0.818	0.285	0.605	-
Res50	288x384	Stereo	gpp	Official	0.280	0.371	0.972	0.335	0.530	Baidu/Google
Res50	288x384	Stereo		Trained	0.261	0.301	0.851	0.299	0.591	Baidu/Google
Res50	288x384	Stereo	gpp	Trained	0.258	0.293	0.841	0.296	0.596	Baidu/Google

• Reported means that the results are reported in the paper.
• Official means that the results are predicted with the models got from their Official Implementations.
• Trained means that the results are predicted with the models trained with this repository.
• code for all the download links of pan Baidu is smde.