Analysis of local binary pattern using uniform bins as palm vein pattern descriptor

Palm vein authentication technology which reads the features of palm vein has been widely used in recent years as it offers high accuracy identification and difficult to be forged or impersonated. This paper demonstrates a palm vein recognition system using uniform Local Binary Pattern (LBP) through...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of physics. Conference series Ročník 1502; číslo 1; s. 12043 - 12050
Hlavní autoři: Atikah Mohd Hayat, Nurul, Mohd Noh, Zarina, Mohamad Yatim, Norhidayah, Ahmad Radzi, Syafeeza
Médium: Journal Article
Jazyk:angličtina
Vydáno: Bristol IOP Publishing 01.03.2020
Témata:
Accuracy
Bins
Datasets
K-nearest neighbors algorithm
Physics
Programming languages
ISSN:1742-6588, 1742-6596
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Palm vein authentication technology which reads the features of palm vein has been widely used in recent years as it offers high accuracy identification and difficult to be forged or impersonated. This paper demonstrates a palm vein recognition system using uniform Local Binary Pattern (LBP) through the Python language and R. Python language was used for contrast enhancement, noise reduction and LBP implementation while R was used for classifying palm vein pattern using K-Nearest Neighbour (KNN) classifier. The palm samples come from two datasets which are from the Chinese Academy of Sciences Institute of Automation (CASIA) and self-dataset. The outcomes were the extracted and classified palm vein pattern based on subjects and their accuracy based on each dataset. The accuracy for all uniform LBP bins and selected uniform LBP bins from self-dataset were 87% and 53% respectively; while for CASIA dataset were 60% and 27% respectively. The results show that the accuracy is higher if all uniform LBP bins are used for the recognition purpose.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1742-6588
1742-6596
DOI:10.1088/1742-6596/1502/1/012043