Two new ways of Discriminant Feature Transform

Author

Jundang, Nattapong ; Ongkittikul, Surachai

Author_Institution

Comput. Eng. Grad. Program, Mahanakorn Univ. of Technol., Bangkok, Thailand

fYear

2014

fDate

14-17 May 2014

Firstpage

1

Lastpage

4

Abstract

In this paper, we present the encoding produces technique for texture recognition. This produces are developed with the purpose to replaced the previous method. Firstly, we encode image informations by VTT (Volume Trace Transform), and trace functionals are chosen by RL (Reinforce Learning). The number of VTT results are extracted by the amount of trace functionals. Secondly, we present P-DFT (Plus-Discriminant Feature Transform) and S-DFT (Square-Discriminant Feature Transform) accumulate result´s VTT and constructs 2-D histogram. Finally, The histogram is compared by chi-square test and in conclusion will present the accuracy of recognition rate of both methods based on brodatz texture database.

Keywords

image recognition; image texture; learning (artificial intelligence); transforms; 2D histogram; P-DFT; RL; S-DFT; VTT; brodatz texture database; chi-square test; encoding; image information; plus-discriminant feature transform; recognition rate; reinforce learning; square-discriminant feature transform; texture recognition; trace functional; volume trace transform; Accuracy; Equations; Histograms; Image texture; Learning (artificial intelligence); Mathematical model; Transforms; reinforcement learning; trace transform;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), 2014 11th International Conference on

Conference_Location

Nakhon Ratchasima

Type

conf

DOI

10.1109/ECTICon.2014.6839816

Filename

6839816