KIT | KIT-Bibliothek | Impressum | Datenschutz

A Multispectral Light Field Dataset and Framework for Light Field Deep Learning

Schambach, Maximilian; Heizmann, Michael

Deep learning undoubtedly has had a huge impact on the computer vision community in recent years. In light field imaging, machine learning-based applications have significantly outperformed their conventional counterparts. Furthermore, multi- and hyperspectral light fields have shown promising results in light field-related applications such as disparity or shape estimation. Yet, a multispectral light field dataset, enabling data-driven approaches, is missing. Therefore, we propose a new synthetic multispectral light field dataset with depth and disparity ground truth. The dataset consists of a training, validation and test dataset, containing light fields of randomly generated scenes, as well as a challenge dataset rendered from hand-crafted scenes enabling detailed performance assessment. Additionally, we present a Python framework for light field deep learning. The goal of this framework is to ensure reproducibility of light field deep learning research and to provide a unified platform to accelerate the development of new architectures. The dataset is made available under . The framework is maintained at .

Verlagsausgabe §
DOI: 10.5445/IR/1000125598
Veröffentlicht am 03.11.2020
DOI: 10.1109/ACCESS.2020.3033056
Zitationen: 2
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Industrielle Informationstechnik (IIIT)
Publikationstyp Zeitschriftenaufsatz
Publikationsjahr 2020
Sprache Englisch
Identifikator ISSN: 2169-3536
KITopen-ID: 1000125598
Erschienen in IEEE access
Verlag Institute of Electrical and Electronics Engineers (IEEE)
Band 8
Seiten 193492–193502
Vorab online veröffentlicht am 22.10.2020
Externe Relationen Forschungsdaten/Software
Schlagwörter Dataset; deep learning; disparity; light field imaging; multispectral imaging
Nachgewiesen in Dimensions
Relationen in KITopen
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page