Romanian Journal of Information Science and Technology (ROMJIST)

An open – access publication

  |  HOME  |   GENERAL INFORMATION  |   ROMJIST ON-LINE  |  KEY INFORMATION FOR AUTHORS  |   COMMITTEES  |  

ROMJIST is a publication of Romanian Academy,
Section for Information Science and Technology

Editor – in – Chief:
Academician Dan Dascalu

Secretariate (office):
Adriana Neagu
Adress for correspondence: romjist@nano-link.net (after 1st of January, 2019)

Editing of the printed version: Mihaela Marian (Publishing House of the Romanian Academy, Bucharest)

Technical editor
of the on-line version:
Lucian Milea (University POLITEHNICA of Bucharest)

Sponsors:
• National Institute for R & D
in Microtechnologies
(IMT Bucharest), www.imt.ro
• Association for Generic
and Industrial Technologies (ASTEGI), www.astegi.ro

ROMJIST Volume 22, No. 3-4, 2019, pp. 272-283, Paper no. 633/2019
 

Kestrilia Rega PRILIANTI, Hendry SETIAWAN, Marcelinus Alfasisurya Setya ADHIWIBAWA, Indah ANITA-SARI, Syaiful ANAM, and Tatas Hardo Panintingjati BROTOSUDARMO
Adaptive Neuro-Fuzzy Inference System for Non-linear Classification Problem of Theobroma cacao Image

ABSTRACT: Fine-flavor cacao is one of the cacao varieties with superior characters, especially the flavor quality. In this research, we develop a rapid and non-destructive method to evaluate cacao seed. Normally, it takes several years to identify whether a cacao seed is among the fine-flavor or bulk varieties. However, since it was known that the colors of the leaves can distinguish fine-flavor cacao from bulk cacao, the image classification method can serve as an alternative means of rapid identification. Furthermore, an Adaptive Neuro-Fuzzy Inference System (ANFIS) is applied to provide an intelligence agent that can automatically learn and identify cacao seed as fine-flavor or bulk. The ANFIS is used because designing a robust non-linear mathematical classification model on color image data is not an easy task. Therefore, the soft computing approach is used to resolve the difficulties. Other soft computing classification methods, i.e. Artificial Neural Network (ANN) and Fuzzy Multilayer Perceptron (Fuzzy-MLP), are also applied in order to evaluate the performance of ANFIS. Our experiment demonstrate that ANFIS model is more robust than the other two methods and could classify the cacao leaves at accuracy rate up to 94% for in-sample and 84% for-out sample prediction.

KEYWORDS: Adaptive Neuro Fuzzy Inference System (ANFIS); cacao; image classification; non-linear classification

Read full text (pdf)






  |  HOME  |   GENERAL INFORMATION  |   ROMJIST ON-LINE  |  KEY INFORMATION FOR AUTHORS  |   COMMITTEES  |