Improve monocular depth estimation without labeled training data by image masking with MIMDepth
Improve monocular depth estimation without labeled training data by image masking with MIMDepth
Image Masking for Robust Self-Supervised Monocular Depth Estimation
arXiv paper abstract https://arxiv.org/abs/2210.02357v1
arXiv PDF paper https://arxiv.org/pdf/2210.02357v1.pdf
Self-supervised monocular depth estimation is a salient task for 3D scene understanding.
… methods have been proposed to predict accurate pixel-wise depth without using labeled data …
Nevertheless, these … focus on … ideal conditions without natural or digital corruptions … absence of occlusions is assumed even for object-specific depth estimation.
… propose MIMDepth, a method that adapts masked image modeling (MIM) for self-supervised monocular depth estimation.
While MIM has been used to learn generalizable features during pre-training … show how it could be adapted for direct training of monocular depth estimation.
… experiments show that MIMDepth is more robust to noise, blur, weather conditions, digital artifacts, occlusions, as well as untargeted and targeted adversarial attacks.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
