Article Per Year (5 Year)
p-Index From 2021 - 2026
1.829
P-Index
This Author published in this journals
All Journal Jurnal Agronomi Indonesia (Indonesian Journal of Agronomy) Jurnal Sains Dasar Jurnal Teknologi Informasi dan Ilmu Komputer Jurnal Pengabdian Masyarakat MIPA dan Pendidikan MIPA Education Achievement: Journal of Science and Research Balance : Jurnal Akuntansi dan Manajemen International Journal of Health Science and Technology JURMADIKTA Jurnal Pengabdian Masyarakat dan Riset Pendidikan Berita Biologi
Annisa Latifa
Universitas Negeri Yogyakarta
Agriculture, Biological Sciences & Forestry Humanities Biochemistry, Genetics & Molecular Biology Computer Science & IT Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Engineering Health Professions Languange, Linguistic, Communication & Media Library & Information Science Mathematics Nursing Public Health Other

Published : 11 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 11 Documents
Search

 1   2 
Implementasi Deep Learning Algoritma Convolutional Neural Network untuk Klasifikasi Kesegaran Buah dan Sayur Latifa, Annisa; Hikmah, Nor; Kurniawan, Hendra; Rohmat Hidayat, Kardilah; Larasati, Niken; Rumini
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 13 No 2: April 2026
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.2026131

Abstract

Buah dan sayur merupakan sumber utama vitamin, mineral, dan serat yang sangat penting untuk menjaga kesehatan tubuh. WHO merekomendasikan konsumsi sebesar 400 gram per hari untuk gizi seimbang. Namun, kualitas dan kesegaran buah dan sayur sering kali sulit diidentifikasikan secara manual, terutama dalam skala besar, karena metode tradisional memiliki keterbatasan akurasi dan rentan terhadap kesalahan manusia. Kemajuan kecerdasan buatan, khususnya deep learning, memberikan solusi inovatif dalam klasifikasi citra. Convolutional Neural Network (CNN), telah terbukti efektif dalam pengenalan dan klasifikasi gambar. Penelitian ini menerapkan CNN dengan arsitektur Inception V3 dalam mengklasifikasikan kesegaran buah dan sayuran menjadi dua kategori utama, yaitu segar dan busuk. Model dikembangkan menggunakan dataset yang terdiri dari 11. 441 citra yang gambar, yang dibagi ke dalam tiga subset utama, yaitu data latih (±44.38%), data validasi (±11.07%), dan data uji (±44.55%). Dengan data kelas terbagi 14 kelas. Hasil penelitian  dengan menggunakan confusion matric  nilai accuracy sebesar 95%  dan hasil evaluasi validation accuracy  sebesar 100% pada epoch ke-4, dengan val_loss terendah sebesar 0.0260  serta nilai MAE  0.26, yang artinya model memiliki kinerja yang sangat baik  dalam mendekteksi kesegaran  buah dan sayur. Penelitian lanjutan disarankan untuk meningkatkan generalisasi model dengan menggunakan dataset yang lebih beragam, dan mengintegrasikan komputasi tepi (edge computing) untuk inspeksi kualitas langsung di Lokasi.   Abstract Fruits and vegetables are primary sources of vitamins, minerals, and fiber, which are essential for maintaining a healthy body. The World Health Organization (WHO) recommends a daily intake of 400 grams for a balanced diet. However, the quality and freshness of fruits and vegetables are often difficult to identify manually, especially at large scale, as traditional methods have limitations in accuracy and are prone to human error. Advances in artificial intelligence, particularly deep learning, offer innovative solutions in image classification. Convolutional Neural Networks (CNNs) have proven effective in image recognition and classification tasks. This study implements a CNN using the Inception V3 architecture to classify the freshness of fruits and vegetables into two main categories: fresh and rotten. The model was developed using a dataset consisting of 11,441 images, divided into three main subsets: training data (approximately 44.38%), validation data (approximately 11.07%), and test data (approximately 44.55%), with 14 distinct classes. The results of the study, based on the confusion matrix, show an accuracy of 95%, and a validation accuracy of 100% at the 4th epoch, with the lowest validation loss recorded at 0.0260 and a MAE of 0.26. These results indicate that the model performs very well in detecting the freshness of fruits and vegetables. Further research is recommended to improve model generalization using more diverse datasets and to integrate edge computing for on-site quality inspection.
Co-Authors Astika Cahyarani Budiwati Budiwati, Budiwati - Diah Rachmawati Dyatmiko, Hendrian Edisah Putra Nainggolan Eko Swi Damarwan Hendra Kurniawan Hikmah, Nor Husmarini, Novi Illahi, Annisasufitri Bryan Karim Karim, Karim Melani, Melyann murdiyanto, joko Niken Larasati Rasyid, Rico Firman RATNAWATI Ratnawati Ratnawati Reflinda, Reflinda Rohmat Hidayat, Kardilah Rumini Salsabila, Huda Sari, Asdini Sekarlangit, Sekarlangit Sudarsono Sudarsono Toto Indriyatmoko Veni Roza Wardhana, Alex Sandria Jaya