| Autor: |
Ghouzali, Sanaa |
| Předmět: |
|
| Zdroj: |
International Review on Computers & Software; Dec2023, Vol. 18 Issue 2, p85-89, 5p |
| Abstrakt: |
The proposed work on the automatic multispectral retinal blood vessel analysis for detecting eye diseases uses the Discrete Wavelet Transform and Local Binary Patterns for feature extraction procedure. The sub-bands of the Discrete Wavelet Transform provide the horizontal, vertical and diagonal details of the image under analysis at different frequency levels. The resulting horizontal, vertical and diagonal sub-bands are then given to the Local Binary Patterns module that generates the local features of the directional wavelet sub-bands. The classification in healthy and glaucomatous images is done using Support Vector Machines. The obtained results show better performance than state of the art algorithms found in the literature for the High-Resolution Fundus image database. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
