SAda-Net: A Self-supervised Adaptive Stereo Estimation CNN For Remote Sensing Image Data

Dominik Hirner*, Friedrich Fraundorfer

*Korrespondierende/r Autor/-in für diese Arbeit

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in einem KonferenzbandBegutachtung

Abstract

Stereo estimation has made many advancements in recent years with the introduction of deep-learning. However the traditional supervised approach to deep-learning requires the creation of accurate and plentiful ground-truth data, which is expensive to create and not available in many situations. This is especially true for remote sensing applications, where there is an excess of available data without proper ground truth. To tackle this problem, we propose a self-supervised CNN with self-improving adaptive abilities. In the first iteration, the created disparity map is inaccurate and noisy. Leveraging the left-right consistency check, we get a sparse but more accurate disparity map which is used as an initial pseudo ground-truth. This pseudo ground-truth is then adapted and updated after every epoch in the training step of the network. We use the sum of inconsistent points in order to track the network convergence. The code for our method will be made available after acceptance at https://github.com/thedodo/SAda-Net.

Originalspracheenglisch
TitelPattern Recognition - 27th International Conference, ICPR 2024, Proceedings
Redakteure/-innenApostolos Antonacopoulos, Subhasis Chaudhuri, Rama Chellappa, Cheng-Lin Liu, Saumik Bhattacharya, Umapada Pal
Herausgeber (Verlag)Springer Science and Business Media Deutschland GmbH
Seiten159-175
Seitenumfang17
ISBN (Print)9783031781919
DOIs
PublikationsstatusVeröffentlicht - 2025
Veranstaltung27th International Conference on Pattern Recognition, ICPR 2024 - Kolkata, Indien
Dauer: 1 Dez. 20245 Dez. 2024

Publikationsreihe

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Band15310 LNCS
ISSN (Print)0302-9743
ISSN (elektronisch)1611-3349

Konferenz

Konferenz27th International Conference on Pattern Recognition, ICPR 2024
Land/GebietIndien
OrtKolkata
Zeitraum1/12/245/12/24

ASJC Scopus subject areas

  • Theoretische Informatik
  • Allgemeine Computerwissenschaft

Fingerprint

Untersuchen Sie die Forschungsthemen von „SAda-Net: A Self-supervised Adaptive Stereo Estimation CNN For Remote Sensing Image Data“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren