All results

The configuration files for all the experiments described below are stored here.

Training loss

Quantitative results

depth loss	grad loss	norm loss	MAE	MSE	RMSE	ABS_REL	LG10	DELTA1	DELTA2	DELTA3
✔	✔	✔	0.3388	0.3150	0.5613	0.1283	0.0548	0.8407	0.9673	0.9907
✔	✔		0.3408	0.3208	0.5664	0.1281	0.0551	0.8381	0.9653	0.9909
✔		✔	0.3519	0.3370	0.5805	0.1311	0.0570	0.8304	0.9622	0.9897
✔			0.3505	0.3408	0.5838	0.1303	0.0567	0.8314	0.9618	0.9891

Other metrics

depth loss	grad loss	norm loss	EDGE_ACCURACY	EDGE_PRECISION	EDGE_RECALL	EDGE_F1SCORE
✔	✔	✔	0.8562	0.6325	0.4913	0.5461
✔	✔		0.8558	0.6360	0.4823	0.5409
✔		✔	0.8474	0.5969	0.4972	0.5334
✔			0.8401	0.5679	0.5020	0.5238

Qualitative results

Each column has the visualizations for the same data.

From top to down...

input images
gt depth visualization
prediction visualization w/ full loss
prediction visualization w/ depth loss + grad loss
prediction visualization w/ depth loss + norm loss
prediction visualization w/ only depth loss