Submitted by
Stephan Richter
Authors
Jin Han Lee, Myung-Kyu Han, Dong Wook Ko, Il Hong Suh
Venue
arxiv
Abstract
From Big to Small: Multi-Scale Local Planar Guidance for Monocular Depth Estimation, retrained with all training data from the Robust Vision Challenge (KITTI, Sintel, rabbitai, VIPER) by rabbitai.
Runtime
88 ms
External data used
KITTI, MPI Sintel, rabbitai
Links
Number of submissions
1

Detailed results
MetricAlldaysunsetsnowrainnight
Scale invariant logarithmic error 1.41 0.96 1.54 1.66 1.31 1.56
Absolute squared error [%] 37.30 16.93 39.08 36.34 27.90 66.26
Root mean squared error of the inverse depth 187.42 89.70 220.79 258.01 148.33 220.28