IMPLEMENTING DATA AUGMENTATION USING DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS TO ENHANCE CNN PERFORMANCE IN WEB-BASED SPICE IMAGE CLASSIFICATION

Daegal, Andre (2025) IMPLEMENTING DATA AUGMENTATION USING DEEP CONVOLUTIONAL GENERATIVE ADVERSARIAL NETWORKS TO ENHANCE CNN PERFORMANCE IN WEB-BASED SPICE IMAGE CLASSIFICATION. Tugas Akhir thesis, University of Technology Yogyakarta.

Text #Abstrak_Andre Daegal_5211811004_Sains Data.pdf Download (159kB)

Abstract

ABSTRACT Adequate data availability is critical in developing effective machine learning models, especially in image classification tasks. This study addresses the challenge of limited training data by applying generative AI–based augmentation using Deep Convolutional Generative Adversarial Networks (DCGAN). The method successfully generated 500 synthetic images for each spice category, doubling the dataset size from 2,000 to 4,000 images. A Convolutional Neural Network (CNN) classification model demonstrated notable improvements, with training accuracy increasing by 8.03% (from 91.27% to 99.30%) and test accuracy rising by 9.60% (from 89.00% to 98.60%). For the ginger class, precision improved from 0.74 to 0.96, with a stable recall and an increase in F1-score from 0.83 to 0.98. The aromatic ginger (kencur) class showed precision growth from 0.91 to 0.99, recall improved to 0.97, and F1-score rose from 0.81 to 0.98. Turmeric retained maximum precision with a slight drop in recall to 0.98 and an F1-score of 0.99. Meanwhile, galangal achieved perfect scores in precision, recall, and F1-score. These findings highlight the significant role of generative AI, particularly DCGAN, in enriching limited datasets and substantially boosting the performance of CNN-based image classification models in web-based applications. Keywords: Augmentation, Data, DCGAN, Image, CNN.

Item Type:	Thesis (Skripsi, Tugas Akhir or Kerja Praktek) (Tugas Akhir)
Subjects:	T Technology > T Technology (General)
Divisions:	Fakultas Sains Dan Teknologi > Data Science
Depositing User:	Sains Data
Date Deposited:	07 Aug 2025 03:44
Last Modified:	07 Aug 2025 03:45
URI:	http://eprints.uty.ac.id/id/eprint/18381

Actions (login required)

View Item