Hermawan, Ramadhan Anugrah and Taufik, Taufik and Gerhana, Yana Aditia (2025) Klasifikasi citra ras kucing berbasis CNN dengan metode MobileNet-V2. INTERNAL (Information System Journal), 8 (1). pp. 70-87. ISSN 2656-0259
|
Text
artikel.pdf Download (903kB) | Preview |
Abstract
Penelitian ini bertujuan untuk mengevaluasi kinerja Convolutional Neural Network dengan arsitektur MobileNet-V2 dalam mengklasifikasikan citra empat ras kucing. Klasifikasi objek merupakan bidang penting dalam computer vision yang berfokus pada informasi visual dari citra digital yang dapat diproses dan dimanipulasi oleh komputer. Permasalahan utama kurangnya pemahaman masyarakat terhadap perbedaan karakteristik fisik setiap ras kucing, yang disebabkan oleh banyaknya kucing hasil perkawinan antar ras dengan kemiripan visual yang sama. Proses pelatihan model dilakukan menggunakan beberapa epoch, dengan setiap epoch diuji sebanyak 10 kali untuk mengamati nilai akurasi pelatihan dan akurasi validasi serta efektivitas dan kestabilan model. Dataset yang digunakan dalam penelitian ini diperoleh dari Kaggle dan situs fiveweb.org dengan total 4.140 citra. Hasil pengujian menunjukkan bahwa epoch 100 menghasilkan akurasi validasi sebesar 94,49%. Meskipun hasil tersebut sudah cukup baik, optimasi diperlukan untuk mengurangi overfitting dan meningkatkan kemampuan generalisasi model. Penelitian ini berkontribusi dalam pengembangan sistem otomatis dapat dimanfaatkan dalam bidang edukasi. This study aims to evaluate the performance of Convolutional Neural Network with MobileNet-V2 architecture in classifying images of four cat breeds. Object classification is an important field in computer vision that focuses on visual information from digital images that can be processed and manipulated by computers. The main problem is the lack of public understanding of the differences in the physical characteristics of each cat breed, which is caused by the large number of cats resulting from interbreeding with similar visual appearances. The model training process was carried out using several epochs, with each epoch tested 10 times to observe the training accuracy and validation accuracy as well as the effectiveness and stability of the model. The dataset used in this study was obtained from Kaggle and the website fiveweb.org, with a total of 4,140 images. The test results showed that epoch 100 produced a validation accuracy of 94.49%. Although this result is quite good, optimization is needed to reduce overfitting and improve the generalization ability of the model. This research contributes to the development of automated systems that can be utilized in the field of education.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | Akurasi; Deep Learning; Epoch; Klasifikasi Kucing; MobileNet-V2 |
| Subjects: | Data Processing, Computer Science Special Computer Methods > Artificial Intelligence Operations, Archieves, Information Centers > Classification of Specific Subject |
| Divisions: | Fakultas Sains dan Teknologi > Program Studi Teknik Informatika |
| Depositing User: | Ramadhan Anugrah Hermawan |
| Date Deposited: | 11 Feb 2026 03:57 |
| Last Modified: | 11 Feb 2026 03:57 |
| URI: | https://digilib.uinsgd.ac.id/id/eprint/128047 |
Actions (login required)
![]() |
View Item |



