V4: MMR algoritmy adaptované pro Latinsku Ameriku

V4: MMR algorithms adapted for Latin America

software

In order to use the result by another entity, it is always necessary to acquire a license

The licensor requires a license fee for the result

Juhas Miroslav, Ing.
Honec Peter, Ing., Ph.D. (UAMT)
Klobušníková Zuzana
Fučík Otto, doc. Dr. Ing. (DCSY)
Šustek Jiří
Ševčík Václav, Ing.
Zemčík Pavel, prof. Dr. Ing., dr. h. c. (DCGM)
Juránek Roman, Ing., Ph.D. (DCGM)
Nosko Svetozár, Ing. (DCGM)

MMR, Model and Make Recognition, Traffic Counter, Convolutional Neural Network, ResNet, High Precision, Brazil, South America

A software module for visual vehicle classification, MMR (Model and Make Recognition), operating with high accuracy based on neural networks (CNNs) and adapted to the appearance of Latin American vehicles thorugh re-training exploiting Latin Americal data set of vehicles. It can be used both as a complement to classification algorithms in WIM systems and as an independent non-intrusive vehicle counter and classifier based on a camera system.

Podléhá obchodnímu tajemství

WIM Latin America, EU, OP PIK - Aplikace, EG20_321/0025242, 2021-2023, running

Computer Graphics Research Group (RG GRAPH)

Department of Computer Graphics and Multimedia (DCGM)