|
Publication
|
Research Title |
LOCAL THREE-VALUED FOUR-DIRECTIONAL PATTERN COUNTING METHOD FOR TEXTURE IMAGE CLASSIFICATION |
Date of Distribution |
25 December 2021 |
Conference |
Title of the Conference |
ISTANBUL International Modern Scientific Research Congress II |
Organiser |
IKSAD Publishing House |
Conference Place |
Online |
Province/State |
Istanbul,Turrkey |
Conference Date |
23 December 2021 |
To |
25 December 2021 |
Proceeding Paper |
Volume |
2021 |
Issue |
2 |
Page |
389 |
Editors/edition/publisher |
IKSAD Publishing House |
Abstract |
Image classification has been widely used nowadays for intelligence vision applications. It comprises a well-designed algorithm that extracts local and global features from images to construct their identities for identification and classification. During the last two decades, local pattern descriptors and many variants of local pattern methods have been proposed for texture,
face, and other image classifications due to their computational simplicity and promising results
in various applications. We introduce a new local three-valued four-directional pattern counting method (LTFPC) for texture image classification. The proposed method is aimed as an alternative local descriptor that is simple and extendible or modifiable to suit general image classification tasks. LTFPC modifies a local binary pattern method (LBP) by considering the directions of subpatterns. It extracts local features of an image by using three comparison values and counting subpatterns in the four main compass directions for constructing the pattern vectors and the histogram of their distribution. The performances of LTFPC are compared with those of LBP and its improvement variants, ILBP and MBLBP, on seven texture databases. The experimental results show that the LTFPC method overall outperforms the compared methods. |
Author |
|
Peer Review Status |
ไม่มีผู้ประเมินอิสระ |
Level of Conference |
นานาชาติ |
Type of Proceeding |
Full paper |
Type of Presentation |
Oral |
Part of thesis |
true |
Presentation awarding |
false |
Attach file |
|
Citation |
0
|
|
|
|
|
|
|