Performance Comparison of CNN Transfer Learning Models for Coffee Bean Quality Classification

  • Nur Muhammad Fadli Politeknik Negeri Jember
  • Prawidya Destarianto Politeknik Negeri Jember
  • Hendra Yufit Riskiawan Politeknik Negeri Jember
  • Bekti Maryuni Susanto Politeknik Negeri Jember
  • Satrio Adi Priyambada National Research and Innovation Agency
  • Wawan Hendriawan Nur National Research and Innovation Agency
  • Mukhamad Angga Gumilang Politeknik Negeri Jember
DOI: https://doi.org/10.25047/jtit.v12i2.457
Keywords: Transfer Learning, Coffee Bean Classification, CNN, Xception, MobileNet V2, EfficientNet B1

Abstract

According to SNI Standard No. 01-2907-2008, accurate sorting of coffee beans is crucial for improving export value. Manual sorting is time-consuming, subjective, and error-prone, especially when visual differences are subtle between roast levels. This study proposes and evaluates an automatic, machine-learning based system to support quality assurance in coffee production. We compare three transfer-learning CNN architectures: Xception, MobileNetV2, and EfficientNet-B1 on a publicly available dataset of 1,600 coffee bean images divided into four classes (dark, medium, light, green). All models were trained with the same preprocessing and hyperparameter settings. EfficientNet-B1 achieved the highest test accuracy (100%), followed by Xception (99.5%) and MobileNetV2 (97%). We discuss trade-offs between accuracy and computational efficiency and recommend EfficientNet-B1 for high-accuracy applications and MobileNetV2 for edge/mobile deployment.

Author Biographies

Nur Muhammad Fadli, Politeknik Negeri Jember

Nur Muhammad Fadli, was born in Bondowoso, East Java in 2003. He is currently pursuing her undergraduate degree in Informatics Engineering at Politeknik Negeri Jember, Indonesia. His interests revolve around mobile development, internet of things and machine learning. As part of her academic research, He is also exploring traceability system for coffee products and disease detection in coffee leaves using machine learning. He can reached via email at nmuhammadfadli48@gmail.com.

 

Prawidya Destarianto, Politeknik Negeri Jember

PRAWIDYA DESTARIANTO, holds a Bachelor's degree in Computer Science from STMIK PPKIA Pradnya Paramita and a Master's degree in Engineering from Brawijaya University. He currently teaches at the Department of Information Technology at Jember State Polytechnic. His research interests include artificial intelligence, the Internet of Things, and information systems. He can be contacted via email at prawidya@polije.ac.id

Hendra Yufit Riskiawan, Politeknik Negeri Jember

HENDRA YUFIT RISKIAWAN, was born in Jember, East Java, Indonesia, in 1983. He received his Bachelor degree from Bogor Agricultural University, Indonesia, in 2005 in Computer Science Education, and his Master's degree from Gadjah Mada University, Indonesia, in 2015, in Computer Science. His research interests include Information Systems, the Internet of Things, and machine learning. He can be contacted via email: yufit@polije.ac.id

Bekti Maryuni Susanto, Politeknik Negeri Jember

BEKTI MARYUNI SUSANTO, was born in Yogyakarta Province, Indonesia, in 1984. He received the Bachelor degree from the Yogyakarta State University, Indonesia in 2010 in Electrical Engineering Education and the Master degree from the STMIK Nusa Mandiri Jakarta, Indonesia, in 2012, in Computer Science. His research interests include cloud computing, internet of things, and machine learning. He can be contacted at email: bekti@polije.ac.id.

Satrio Adi Priyambada, National Research and Innovation Agency

SATRIO ADI PRIYAMBADA, holds a Bachelor's  degree in Information Systems from the Sepuluh Nopember Institute of Technology (ITS), a Master's degree in Management Information Systems from the Sepuluh Nopember Institute of Technology (ITS), and a Doctor of Philosophy - PhD in Human and Environmental Informatics from Kumamoto University. He is currently a Researcher at the Indonesian National Research and Innovation Agency (BRIN RI). His research interests include Process Mining, Data Mining, and Machine Learning. He can be contacted via email at satr014@brin.go.id.

Wawan Hendriawan Nur, National Research and Innovation Agency

WAWAN HENDRIAWAN NUR, holds a Bachelor’s of Science degree from Padjadjaran University and a Master of Informatics degree from the Bandung Institute of Technology (ITB). He currently serves as the Geographic Information System (GIS) Administrator at the National Research and Innovation Agency (BRIN RI). His research interests include Geographic Information Systems (GIS) and Spatial Information Systems. He can be contacted via email at wawa020@brin.go.id.

Mukhamad Angga Gumilang, Politeknik Negeri Jember

MUKHAMAD ANGGA GUMILANG, holds a Bachelor in Informatics Engineering Education from State University of Malang, Master of Engineering (M. Eng.) from Gadjah Mada University. He is currently lecturing in Information Techology department of State Polytechnic Jember. His research areas in artificial intelligence, social media analytics and human- computer interaction. He can be contacted at email: angga.gumilang@polije.ac.id.

Published
2025-12-31
Issue
Vol 12 No 2 (2025): December
Section
Articles
Copyright (c) 2025 Jurnal Teknologi Informasi dan Terapan

The Authors submitting a manuscript do so on the understanding that if accepted for publication, copyright of the article shall be assigned to Jurnal Teknologi Informasi dan Terapan (J-TIT) and Department of Information Technology, Politeknik Negeri Jember as publisher of the journal. Copyright encompasses rights to reproduce and deliver the article in all form and media, including reprints, photographs, microfilms, and any other similar reproductions, as well as translations. Authors should sign a copyright transfer agreement when they have approved the final proofs sent by Jurnal Teknologi Informasi dan Terapan (J-TIT) prior to the publication. The copyright transfer agreement can be download here .

Jurnal Teknologi Informasi dan Terapan (J-TIT) and Department of Information Technology, Politeknik Negeri Jember and the Editors make every effort to ensure that no wrong or misleading data, opinions or statements be published in the journal. In any way, the contents of the articles and advertisements published in Jurnal Teknologi Informasi dan Terapan (J-TIT) are the sole responsibility of their respective authors and advertisers.

Users of this website will be licensed to use materials from this website following the Creative Commons Attribution 4.0 International License. No fees charged. Please use the materials accordingly.

 

Creative Commons License
 This work is licensed under a Creative Commons Attribution-Share A like 4.0 International License

You are free to:

  • Adapt — remix, transform, and build upon the material for any purpose, even commercially.
  • The licensor cannot revoke these freedoms as long as you follow the license terms.