print

Morphological cross entropy loss for improved semantic segmentation of small and thin objects

author
Pihlak, Rene
Riid, Andri
statement of authorship
René Pihlak, Andri Riid
source
Procedia computer science
publisher
Elsevier
journal volume number month
vol. 192
year of publication
2021
pages
p. 582–591
series
Procedia computer science ; 181
conference name, date
25th KES International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, KES 2021, 8 - 10 Sept. 2021
conference location
Szczecin
url
https://doi.org/10.1016/j.procs.2021.08.060
subject term
liiklusmärgid
ekspertsüsteemid
keyword
semantic segmentation
traffic signs
loss function
convolutional neural network (CNN)
ISSN
1877-0509
notes
Bibliogr.: 26 ref
Open Access
Open Access
scientific publication
teaduspublikatsioon
classifier
3.1
Scopus
Conference proceeding at Scopus
Article at Scopus
WOS
Article at WOS
category (general)
Computer science
category (sub)
Computer science. General computer science
kvartiil
Q2
TTÜ department
tarkvarateaduse instituut
language
inglise
Uurimisrühm
Laboratory of proactive technologies
Pihlak, R., Riid, A. Morphological cross entropy loss for improved semantic segmentation of small and thin objects // Procedia computer science (2021) vol. 192, p. 582–591. https://doi.org/10.1016/j.procs.2021.08.060