Logo des Repositoriums
 

Build Your Own Training Data - Synthetic Data for Object Detection in Aerial Images

dc.contributor.authorLaux, Lea
dc.contributor.authorSchirmer, Sebastian
dc.contributor.authorSchopferer, Simon
dc.contributor.authorDauer, Johann
dc.contributor.editorMichael, Judith
dc.contributor.editorPfeiffer, Jérôme
dc.contributor.editorWortmann, Andreas
dc.date.accessioned2022-02-21T05:04:44Z
dc.date.available2022-02-21T05:04:44Z
dc.date.issued2022
dc.description.abstractMachine learning has become one of the most widely used techniques in artificial intelligence, especially for image processing. One of the biggest challenges in developing an accurate image processing model is to collect large amounts of data that are suffi ciently close to the real-world scenario. Ideally, real-world data is therefore used for model training. Unfortunately, real-world data is often insuffi ciently available and expensive to generate. Therefore, models are trained using synthetic data. However, there is no standardized method of how training data is generated and which properties determine the data quality. In this paper, we present fi rst steps towards the generation of large amounts of data for human detection based on aerial images. To create labeled aerial images, we are using Unreal Engine and AirSim. We report on fi rst impressions of the generated labeled aerial images and identify future challenges – current simulation tools can be used to create realistic and diverse images including labeling, but native support would be benefi cial to ease their usage.en
dc.identifier.doi10.18420/se2022-ws-18
dc.identifier.urihttps://dl.gi.de/handle/20.500.12116/38363
dc.language.isoen
dc.publisherGesellschaft für Informatik e.V.
dc.relation.ispartofSoftware Engineering 2022 Workshops
dc.relation.ispartofseriesKeine
dc.subjectMachine Learning
dc.subjectSynthetic Data
dc.subjectSimulation Environment
dc.subjectUnmanned Aircraft
dc.subjectHuman Detection
dc.titleBuild Your Own Training Data - Synthetic Data for Object Detection in Aerial Imagesen
dc.typeText/Conference Paper
gi.citation.endPage190
gi.citation.publisherPlaceBonn
gi.citation.startPage182
gi.conference.date21.- 25. Februar
gi.conference.locationBerlin (virtuell)
gi.conference.sessiontitleAvioSE

Dateien

Originalbündel
1 - 1 von 1
Lade...
Vorschaubild
Name:
paper14.pdf
Größe:
16.61 MB
Format:
Adobe Portable Document Format