Deteksi Klasifikasi Gambar Makanan Nabati untuk Persyaratan Sertifikasi Halal

Jasir, Abdullah Nasih (2025) Deteksi Klasifikasi Gambar Makanan Nabati untuk Persyaratan Sertifikasi Halal. Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5025211111-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (666kB) | Request a copy

Abstract

Penelitian ini bertujuan untuk mengembangkan sistem klasifikasi gambar makanan nabati yang berkaitan dengan kebutuhan sertifikasi halal. Sistem ini dibangun menggunakan beberapa arsitektur pre-trained, seperti MobileNet, Inception, ResNet, dan EfficientNet. Pengujian dilakukan dalam tiga skenario, yaitu penggunaan base model, fine-tuning, dan hyperparameter tuning. Data set berisi gambar makanan nabati yang telah dikategorikan berdasarkan kebutuhan sertifikasi halal. Hasil menunjukkan bahwa MobileNet dan Inception memberikan akurasi terbaik, terutama setelah melalui proses fine-tuning. Sebaliknya, ResNet dan EfficientNet menunjukkan performa yang kurang optimal tanpa
penyesuaian lebih lanjut. Temuan ini menunjukkan bahwa pemilihan arsitektur dan metode pelatihan berpengaruh signifikan terhadap akurasi model.
======================================================================================================================================
This study aims to develop a plant-based food image classification system relevant to halal certification requirements. The system was built using several pre-trained architectures, including MobileNet, Inception, ResNet, and EfficientNet. Testing was conducted in three scenarios: base model use, fine-tuning, and hyperparameter tuning. The dataset contained images of plant-based foods categorized based on their halal certification requirements. The results showed that MobileNet and Inception provided the best accuracy, especially after fine-tuning. Conversely, ResNet and EfficientNet performed less than optimally without further adjustment. These findings indicate that the choice of architecture and training method significantly influences model accuracy

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Fine-Tuned, Halal, Hyperparameter Tuning, Klasifikasi Gambar, Transformer, Image Classification
Subjects:	T Technology > T Technology (General)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Abdullah Nasih Jasir
Date Deposited:	14 Jul 2025 04:09
Last Modified:	14 Jul 2025 04:09
URI:	http://repository.its.ac.id/id/eprint/119676

Actions (login required)

View Item