KIT | KIT-Bibliothek | Impressum | Datenschutz

People Detection in a Depth Sensor Network via Multi-View CNNs trained on Synthetic Data

Wetzel, Johannes; Zeitvogel, Samuel; Laubenheimer, Astrid; Heizmann, Michael

Abstract (englisch):
In this work an approach for wide-area indoor people detection with a network of depth sensors is presented. We propose an end-to-end multi-view deep learning architecture which takes three foreground segmented overlapping depth images as input and predicts the marginal probability distribution of people present in the scene. In contrast to classical data-driven approaches our method does not make use of any real image data for training but uses a randomized generative scene model to generate synthetic depth images which are used to train our proposed deep learning architecture. The evaluation shows promising results on a publicly available data set.



Originalveröffentlichung
DOI: 10.1109/ISETC50328.2020.9301076
Zugehörige Institution(en) am KIT Institut für Industrielle Informationstechnik (IIIT)
Publikationstyp Proceedingsbeitrag
Publikationsdatum 05.11.2020
Sprache Englisch
Identifikator ISBN: 978-1-7281-8921-5
KITopen-ID: 1000128148
Erschienen in 2020 International Symposium on Electronics and Telecommunications (ISETC), 020 International Symposium on Electronics and Telecommunications (ISETC), Timisoara, Romania Fundin, 5-6 Nov. 2020
Verlag Institute of Electrical and Electronics Engineers (IEEE)
Seiten 1–4
Schlagwörter multi-view person detection; network of depth cameras; top-view people detection; synthetic depth images; multi-view CNN architecture
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page