Does Monocular Depth Estimation Provide Better Pre-training than Classification for Semantic Segmentation?

Dong Lao Alex Wong Stefano Soatto

Published in: CoRR (2022)

Keyphrases

street scenes
semantic segmentation
depth estimation
training set
dynamic scenes
machine learning
stereo matching
feature space
training examples
depth map
pose estimation
image matching
conditional random fields
higher order
dynamic programming
feature extraction
image sequences