Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles

Andreas Kamilaris; Corjan van den Brink; Savvas Karatsiolis

doi:10.48550/arXiv.1908.06472

Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles

Andreas Kamilaris, Corjan van den Brink, Savvas Karatsiolis

Research output: Working paper › Preprint › Academic

9 Downloads (Pure)

Abstract

This paper describes preliminary work in the recent promising approach of generating synthetic training data for facilitating the learning procedure of deep learning (DL) models, with a focus on aerial photos produced by unmanned aerial vehicles (UAV). The general concept and methodology are described, and preliminary results are presented, based on a classification problem of fire identification in forests as well as a counting problem of estimating number of houses in urban areas. The proposed technique constitutes a new possibility for the DL community, especially related to UAV-based imagery analysis, with much potential, promising results, and unexplored ground for further research.

Original language	English
Publisher	ArXiv.org
DOIs	https://doi.org/10.48550/arXiv.1908.06472
Publication status	Published - 18 Aug 2019

Keywords

cs.CV
cs.LG
eess.IV

Access to Document

10.48550/arXiv.1908.06472Licence: CC BY

ArticleFinal published version, 1.29 MBLicence: CC BY

1 Conference contribution

Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles
Kamilaris, A., van den Brink, C. & Karatsiolis, S., 23 Aug 2019, Computer Analysis of Images and Patterns : CAIP 2019 International Workshops, ViMaBi and DL-UAV, Salerno, Italy, September 6, 2019, Proceedings. Vento, M., Percannella, G., Colantonio, S., Giorgi, D., Matuszewski, B. J., Kerdegari, H. & Razaak, M. (eds.). Cham: Springer, p. 81-90 (Communications in Computer and Information Science; vol. 1089).
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Open Access
File
8 Citations (Scopus)

95 Downloads (Pure)

Cite this

@techreport{1aeb1df037544e22aeb6fde29a1383c9,

title = "Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles",

abstract = "This paper describes preliminary work in the recent promising approach of generating synthetic training data for facilitating the learning procedure of deep learning (DL) models, with a focus on aerial photos produced by unmanned aerial vehicles (UAV). The general concept and methodology are described, and preliminary results are presented, based on a classification problem of fire identification in forests as well as a counting problem of estimating number of houses in urban areas. The proposed technique constitutes a new possibility for the DL community, especially related to UAV-based imagery analysis, with much potential, promising results, and unexplored ground for further research. ",

keywords = "cs.CV, cs.LG, eess.IV",

author = "Andreas Kamilaris and {van den Brink}, Corjan and Savvas Karatsiolis",

note = "Workshop on Deep-learning based computer vision for UAV in conjunction with CAIP 2019, Salerno, italy, September 2019",

year = "2019",

month = aug,

day = "18",

doi = "10.48550/arXiv.1908.06472",

language = "English",

publisher = "ArXiv.org",

type = "WorkingPaper",

institution = "ArXiv.org",

}

TY - UNPB

T1 - Training Deep Learning Models via Synthetic Data

T2 - Application in Unmanned Aerial Vehicles

AU - Kamilaris, Andreas

AU - van den Brink, Corjan

AU - Karatsiolis, Savvas

N1 - Workshop on Deep-learning based computer vision for UAV in conjunction with CAIP 2019, Salerno, italy, September 2019

PY - 2019/8/18

Y1 - 2019/8/18

N2 - This paper describes preliminary work in the recent promising approach of generating synthetic training data for facilitating the learning procedure of deep learning (DL) models, with a focus on aerial photos produced by unmanned aerial vehicles (UAV). The general concept and methodology are described, and preliminary results are presented, based on a classification problem of fire identification in forests as well as a counting problem of estimating number of houses in urban areas. The proposed technique constitutes a new possibility for the DL community, especially related to UAV-based imagery analysis, with much potential, promising results, and unexplored ground for further research.

AB - This paper describes preliminary work in the recent promising approach of generating synthetic training data for facilitating the learning procedure of deep learning (DL) models, with a focus on aerial photos produced by unmanned aerial vehicles (UAV). The general concept and methodology are described, and preliminary results are presented, based on a classification problem of fire identification in forests as well as a counting problem of estimating number of houses in urban areas. The proposed technique constitutes a new possibility for the DL community, especially related to UAV-based imagery analysis, with much potential, promising results, and unexplored ground for further research.

KW - cs.CV

KW - cs.LG

KW - eess.IV

U2 - 10.48550/arXiv.1908.06472

DO - 10.48550/arXiv.1908.06472

M3 - Preprint

BT - Training Deep Learning Models via Synthetic Data

PB - ArXiv.org

ER -

Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles

Abstract

Keywords

Access to Document

Fingerprint

Research output

Training Deep Learning Models via Synthetic Data: Application in Unmanned Aerial Vehicles

Cite this