Szczegóły publikacji
Opis bibliograficzny
Performance of LiDAR object detection deep learning architectures based on artificially generated point cloud data from CARLA simulator / Daniel DWORAK, Filip Ciepiela, Jakub Derbisz, Izzat Izzat, Mateusz Komorkiewicz, Mateusz Wójcik // W: MMAR 2019 : 24th international conference on Methods and Models in Automation and Robotics : 26–29 August 2019, Międzyzdroje, Polska : abstracts. — Szczecin : ZAPOL Sobczyk, [2019]. — Dod. e-ISBN 978-1-7281-0933-6. — ISBN: 978-83-7518-922-3; e-ISBN: 978-1-7281-0932-9. — S. 72. — Pełny tekst w: MMAR 2019 [Dokument elektroniczny] : 24th international conference on Methods and Models in Automation & Robotics : August 26–29, 2019, Międzyzdroje, Poland / Faculty of Electrical Engineering. West Pomeranian University of Technology Szczecin. — [Piscataway] : IEEE, cop. 2019. — Dysk Flash. — S. 600–605. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 605, Abstr. — Toż pod adresem https://ieeexplore-1ieee-1org-1000047po00a9.wbg2.bg.agh.edu.pl/stamp/stamp.jsp?tp=&arnumber=8864642
Autorzy (6)
- AGHDworak Daniel
- Ciepiela Filip
- Derbisz Jakub
- Izzat Izzat
- Komorkiewicz Mateusz
- Wójcik Mateusz
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 123699 |
|---|---|
| Data dodania do BaDAP | 2019-09-05 |
| DOI | 10.1109/MMAR.2019.8864642 |
| Rok publikacji | 2019 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
| Konferencja | International Conference on Methods and Models in Automation and Robotics 2019 |
Abstract
Training deep neural network algorithms for LiDAR based object detection for autonomous cars requires huge amount of labeled data. Both data collection and labeling requires a lot of effort, money and time. Therefore, the use of simulation software for virtual data generation environments is gaining wide interest from both researchers and engineers. The big question remains how well artificially generated data resembles the data gathered by real sensors and how the differences affects the final algorithms performance. The article is trying to make a quantitative answer to the above question. Selected state-of-the-art algorithms for LiDAR point cloud object detection were trained on both real and artificially generated data sets. Their performance on different test sets were evaluated. The main focus was to determinate how well artificially trained networks perform on real data and if combined train sets can achieve better results overall. © 2019 IEEE.
