Publication Details

Detekce chodců z výšky pomocí neuronových sítí

KANICH Ondřej, DRAHANSKÝ Martin and GOLDMANN Tomáš. Detekce chodců z výšky pomocí neuronových sítí. Brno, 2019.
English title
Pedestrian detection from height using neural nets
Type
technical report
Language
czech
Authors
Kanich Ondřej, Ing., Ph.D. (DITS FIT BUT)
Drahanský Martin, prof. Ing., Dipl.-Ing., Ph.D. (DITS FIT BUT)
Goldmann Tomáš, Ing. (DITS FIT BUT)
Abstract

Report is dealing with description of methods for detection of person in image. Also it monitors and visualises movement of these persons. Report is focusing on datasets acquired from height, e.g. from drone. For detection RetinaNet neural net model is used. Each detected person gets and feature vector. Using this vector the movement of the persons is monitored. Success rate of detection is 58.6 %.

Published
2019
Pages
38
Place
Brno, CZ
BibTeX 
@TECHREPORT{FITPUB12162,
   author = "Ond\v{r}ej Kanich and Martin Drahansk\'{y} and Tom\'{a}\v{s} Goldmann",
   title = "Detekce chodc\r{u} z v\'{y}\v{s}ky pomoc\'{i} neuronov\'{y}ch s\'{i}t\'{i}",
   pages = 38,
   year = 2019,
   location = "Brno, CZ",
   language = "czech",
   url = "https://www.fit.vut.cz/research/publication/12162"
}
Files
Back to top