Multi-object tracking with deep learning ensemble for unmanned aerial system applications
Xie W, Ide J, Izadi D, Banger S, Walker T, Ceresani R, Spagnuolo D, Guagliano C, Diaz H, Twedt J. Multi-object tracking with deep learning ensemble for unmanned aerial system applications. Proceedings Of SPIE--the International Society For Optical Engineering 2021, 11870: 118700i-118700i-13. DOI: 10.1117/12.2600209.Peer-Reviewed Original ResearchMulti-object trackingUnmanned aerial systemsConvolutional neural network (CNN) encoderUnmanned Aerial System (UAS) applicationsNeural network encoderMultiple hypothesis tracking (MHT) frameworkObject tracking modelReal-time situationsDifferent similarity measuresImage embeddingEntity trajectoriesDeep learningSiamese networkNetwork encoderObject trajectoriesZoom levelAttention mechanismDynamic backgroundObject detectorTracking frameworkMOT methodsIllumination changesLatent spaceSituational awarenessSimilarity measure