Synthetic Data Quantity in Animal Classification

Fieter Brain Pasaribu; Luh Joni Erawati Dewi; I Nyoman Saputra Wahyu Wijaya; Ni Wayan Marti; Ni Putu Novita Puspa Dewi

doi:10.24036/jtip.v20i1.1137

Authors

Fieter Brain Pasaribu Universitas Pendidikan Ganesha
Luh Joni Erawati Dewi Universitas Pendidikan Ganesha
I Nyoman Saputra Wahyu Wijaya Universitas Pendidikan Ganesha
Ni Wayan Marti Universitas Pendidikan Ganesha
Ni Putu Novita Puspa Dewi Universitas Pendidikan Ganesha

DOI:

https://doi.org/10.24036/jtip.v20i1.1137

Keywords:

DCGAN, Image Classification, Synthetic Data, EfficientNetB0, MobileNetV2

Abstract

The need for large datasets remains a challenge in deep learning-based classification models such as Convolutional Neural Networks. Limited data can lead to overfitting, where the model performs well on training data but poorly on testing data. To address this, numerous studies have been conducted using both traditional augmented data and synthetic data generated by Generative Adversarial Networks (GANs). GANs have proven effective across a wide range of dataset domains. However, no research has focused on the effect of the amount of synthetic data on model classification accuracy, leading to confusion about how much should be used to avoid overfitting. Therefore, this study will review the effectiveness of GANs by examining the effect of synthetic data ranging from small amounts to 20 times the original dataset size. The synthetic image results from GANs were then tested using them as supplementary datasets for MobileNetV2 and EfficientNetB0. The results show that the effectiveness of synthetic data fluctuates significantly. At certain amounts, the use of synthetic data can improve accuracy, such as in MobileNetV2, which increased from 73% to 77%. However, at certain amounts, the use of synthetic data also decreases accuracy from 73% as the baseline to 66%. This fluctuating condition is caused by the synthetic data from GANs not being consistent in representing the original data well. The lack of quality of GAN is caused by the architecture used, which is still too simple to capture data patterns with a wide variety of perspectives

References

S. Dong, P. Wang, and K. Abbas, “A survey on deep learning and its applications,” Comput. Sci. Rev., vol. 40, p. 100379, 2021, doi: 10.1016/j.cosrev.2021.100379.

M. Saleem, N. Senan, F. Wahid, M. Aamir, A. Samad, and M. Khan, “Comparative Analysis of Recent Architecture of Convolutional Neural Network,” Math. Probl. Eng., 2022, doi: 10.1155/2022/7313612.

N. L. P. Kurniawati, M. W. A. Kesiman, and I. M. G. Sunarya, “Recognition of Balinese Traditional Ornament Carving Images with Convolutional Neural Network and Discrete Wavelet Transform,” J. Ilm. Tek. Elektro Komput. Dan Inf., vol. 8, no. 4, pp. 670–678, 2023.

K. Y. Mahendra, L. J. E. Dewi, I. K. Purnamawan, and F. B. Pasaribu, “Songket Motif Classification using MobileNet Architecture for Cultural Heritage Preservation,” Int. J. Informatics Comput., vol. 7, no. 2, pp. 725–734, 2025.

A. T. Hidayat, I. M. G. Sunarya, I. P. G. S. Andisana, N. N. E. Smrti, and A. A. G. W. S. Erlangga, “PERBANDINGAN PERFORMA MODEL CNN PADA KLASIFIKASI BIJI KOPI DENGAN CITRA BERBASIS CONDITIONAL LIGHTWEIGHT GAN,” J. Pendidik. Teknol. dan Kejuru., vol. 23, no. 1, pp. 105–116, 2026.

I. M. A. M. Dinata, I. G. A. Gunadi, and I. M. G. Sunarya, “Analisis Hyperparameter Pada Klasifikasi Jenis Daging Menggunakan Algoritma Convolutional Neural Network,” J-SAKTI (Jurnal Sains Komput. dan Inform., vol. 8, no. 1, pp. 25–34, 2024.

I. G. L. T. Sumantara, M. W. A. Kesiman, and I. M. G. Sunarya, “Comparative Analysis of CNN Methods for Periapical Radiograph Classification,” J. Nas. Pendidik. Tek. Inform. JANAPATI, vol. 13, no. 2, pp. 204–214, 2024.

I. D. Maysanjaya and M. Dendi, “Klasifikasi Pneumonia pada Citra X-rays Paru-paru dengan Convolutional neural network,” J. Nas. Tek. Elektro dan Teknol. Inf, vol. 9, no. 2, pp. 190–195, 2020.

I. P. A. Suputra, I. G. A. Gunadi, and I. M. G. Sunarya, “Hyperparameter Optimization with MobileNet Architecture and VGG Architecture for Urban Traffic Density Classification Using Bali Camera Image Data,” Sink. J. dan Penelit. Tek. Inform., vol. 9, no. 3, pp. 1132–1145, 2025.

R. Firmansyah, “Deep learning implementation using convolutional neural network for interest classification,” Faculty of Science and Technology UIN Syarif Hidayatullah Jakarta, 2021.

A. A. Praramadhan and G. E. Saputra, “Cycle generative adversarial networks algorithm with style transfer for image generation,” arXiv Prepr. arXiv2101.03921, 2021.

I. Goodfellow et al., “Generative adversarial nets,” 2014.

Z. Ding, S. Jiang, and J. Zhao, “Take a close look at mode collapse and vanishing gradient in GAN,” 2022, IEEE.

A. Radford, L. Metz, and S. Chintala, “Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks,” arXiv Prepr. arXiv1511.06434, 2015.

B. Liu, J. Lv, X. Fan, J. Luo, and T. Zou, “Application of an improved DCGAN for image generation,” Mob. Inf. Syst., vol. 2022, no. 1, 2022.

I. Papastratis, “Ablation study of self-supervised learning for image classification,” pp. 1–5, 2021, [Online]. Available: http://arxiv.org/abs/2112.02297

J. J. Bird, C. M. Barnes, L. J. Manso, A. Ekárt, and D. R. Faria, “Fruit quality and defect image classification with conditional GAN data augmentation,” Sci. Hortic. (Amsterdam)., vol. 293, p. 110684, 2022.

N. Azahro Choirunisa, T. Karlita, and R. Asmara, “Deteksi Ras Kucing Menggunakan Compound Model Scaling Convolutional Neural Network,” Technomedia J., vol. 6, no. 2, pp. 236–251, 2021, doi: 10.33050/tmj.v6i2.1704.

R. M. Thaniket, M. H. J. Rahanra, and W. W. Dharsono, “Klasifikasi Jenis Burung Cendrawasih Menggunakan Convolutional Neural Network (CNN) Berdasarkan Citra,” J. Ilm. Sist. Inf. dan Tek. Inform., vol. 8, no. 1, pp. 103–113, 2025.

F. B. Pasaribu, L. J. E. Dewi, K. Y. E. Aryanto, and P. Varnakovida, “Effect of Synthetic Data Augmentation on Plant Classification Accuracy Using MobileNetV2, EfficientNet-B0, and ResNet-18,” 2025 IEEE 5th Int. Conf. Softw. Eng. Artif. Intell. SEAI 2025, pp. 1–6, 2025, doi: 10.1109/SEAI65851.2025.11108883.

K. R. M. Manikam, L. J. E. Dewi, K. Y. E. Aryanto, K. A. Seputra, and P. Varnakovida, “Analisis Hyperparameter Pada Klasifikasi Jenis Tanaman Menggunakan Algoritma ResNet50 Dan MobileNetV2,” JATI (Jurnal Mhs. Tek. Inform., vol. 9, no. 6, pp. 9921–9928, 2025.

K. A. Tamayasa and L. J. E. Dewi, “ANALISIS PERBANDINGAN MODEL ARSITEKTUR MOBILENETV2 DAN EFFICIENTNETB3 DALAM KLASIFIKASI PENYAKIT DAUN JAGUNG,” J. Inform. dan Tek. Elektro Terap., vol. 14, no. 1, 2026.

M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L. C. Chen, “Mobilenetv2: Inverted residuals and linear bottlenecks,” 2018.

M. Tan and Q. Le, “Efficientnet: Rethinking model scaling for convolutional neural networks,” 2019, PMLR.

M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, and S. Hochreiter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” 2017.

T. Salimans, I. Goodfellow, W. Zaremba, V. Cheung, A. Radford, and X. Chen, “Improved techniques for training gans,” 2016.