• Home
  • Search
  • Learning Structure-from-Motion from Motion
  • Cite Icon18
  • https://doi.org/10.1007/978-3-030-11015-4_27Copy DOI Icon

Learning Structure-from-Motion from Motion

  • Jan 1, 2019
  • Clément Pinard +3 more
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

This work is based on a questioning of the quality metrics used by deep neural networks performing depth prediction from a single image, and then of the usability of recently published works on unsupervised learning of depth from videos. These works are all predicting depth from a single image, thus it is only known up to an undetermined scale factor, which is not sufficient for practical use cases that need an absolute depth map, i.e. the determination of the scaling factor. To overcome these limitations, we propose to learn in the same unsupervised manner a depth map inference system from monocular videos that takes a pair of images as input. This algorithm actually learns structure-from-motion from motion, and not only structure from context appearance. The scale factor issue is explicitly treated, and the absolute depth map can be estimated from camera displacement magnitude, which can be easily measured from cheap external sensors. Our solution is also much more robust with respect to domain variation and adaptation via fine tuning, because it does not rely entirely on depth from context. Two use cases are considered, unstabilized moving camera videos, and stabilized ones. This choice is motivated by the UAV (for Unmanned Aerial Vehicle) use case that generally provides reliable orientation measurement. We provide a set of experiments showing that, used in real conditions where only speed can be known, our network outperforms competitors for most depth quality measures. Results are given on the well known KITTI dataset [5], which provides robust stabilization for our second use case, but also contains moving scenes which are very typical of the in-car road context. We then present results on a synthetic dataset that we believe to be more representative of typical UAV scenes. Lastly, we present two domain adaptation use cases showing superior robustness of our method compared to single view depth algorithms, which indicates that it is better suited for highly variable visual contexts.

Similar Papers
  • Research Article
  • Citations12

Unsupervised Monocular Depth Perception: Focusing on Moving Objects

  • Dec 15, 2021
  • IEEE Sensors Journal
  • Hualie Jiang +3
  • Research Article
  • Citations12

Depth-Based 3D Face Reconstruction and Pose Estimation Using Shape-Preserving Domain Adaptation

  • Sep 23, 2020
  • IEEE Transactions on Biometrics, Behavior, and Identity Science
  • Yicheng Zhong +8
  • Conference Article
  • Citations13

Dual Adversarial Network for Unsupervised Ground/Satellite-to-Aerial Scene Adaptation

  • Oct 12, 2020
  • Jianzhe Lin +4
  • Research Article
  • Citations19

A new image-quality evaluating and enhancing methodology for bridge inspection using an unmanned aerial vehicle

  • Feb 01, 2021
  • Smart Structures and Systems
  • Hyung‐Jo Jung +5
  • Research Article

Vision-Based UAV 3D Tracking via Relative Pose Estimation

  • Jan 01, 2025
  • IEEE Sensors Journal
  • Junfan Yi +5
  • Research Article
  • Citations80

Sample-to-sample correspondence for unsupervised domain adaptation

  • May 25, 2018
  • Engineering Applications of Artificial Intelligence
  • Debasmit Das +1
  • Research Article
  • Citations88

Unsupervised Domain Adaptation for Depth Prediction from Images.

  • Sep 12, 2019
  • IEEE Transactions on Pattern Analysis and Machine Intelligence
  • Alessio Tonioni +3
  • Research Article
  • Citations17

ParaUDA: Invariant Feature Learning With Auxiliary Synthetic Samples for Unsupervised Domain Adaptation

  • Nov 01, 2022
  • IEEE Transactions on Intelligent Transportation Systems
  • Wenwen Zhang +3
  • Research Article
  • Citations67

Survey on rain removal from videos or a single image

  • Dec 27, 2021
  • Science China Information Sciences
  • Hong Wang +4
  • Conference Article
  • Citations29

Floor detection based depth estimation from a single indoor scene

  • Sep 01, 2013
  • Changhwan Chun +3
  • Conference Article
  • Citations49

Automatic 3-D depth recovery from a single urban-scene image

  • Nov 01, 2012
  • Chen-Yu Tseng +1
  • Research Article
  • Citations7

Geometry understanding from autonomous driving scenarios based on feature refinement

  • Jul 22, 2020
  • Neural Computing and Applications
  • Mingliang Zhai +1
  • Conference Article
  • Citations5

Integration of monocular cues to create depth effect

  • Apr 21, 1997
  • C.T Swain
  • Research Article
  • Citations3

Encoder-Decoder Structure with Multiscale Receptive Field Block for Unsupervised Depth Estimation from Monocular Video

  • Jun 17, 2022
  • Remote Sensing
  • Songnan Chen +4
  • Research Article
  • Citations1

Test‐Time Optimization for Video Depth Estimation Using Pseudo Reference Depth

  • Jan 09, 2023
  • Computer Graphics Forum
  • Libing Zeng +1
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.