Mukti, Muhammad Hari (2025) Penerapan Algoritma Convolutional Neural Network (CNN) untuk Klasifikasi Gambar Motif Batik Tabir Riau Rani. Other thesis, Universitas Islam Riau.

Text skripsi_183510525_watermark.pdf - Published Version Restricted to Registered users only Download (7MB) | Request a copy

Abstract

Batik Tabir Riau Rani is a local cultural heritage featuring distinctive and meaningful motifs, yet it remains largely undocumented through modern technological means. This study leverages Convolutional Neural Network (CNN) algorithms to intelligently classify images of batik motifs. Data collection was a crucial initial step, carried out by capturing primary data directly from batik shops using photography techniques to ensure high-quality and analyzable inputs for CNN processing. The motif documentation involved field visits to batik shops, where 17 distinct batik motifs were photographed from various angles using mobile phone cameras, resulting in 5 to 15 images per motif and an initial total of 108 images. To enhance data volume and variability, image augmentation techniques were applied, producing a dataset of 2,268 images for model training. This study evaluated three CNN architectures: MobileNetV2, DenseNet121, and NASNetMobile. Among them, MobileNetV2 demonstrated the best performance, achieving an accuracy of 98%. These findings indicate that CNN-based approaches not only excel in recognizing batik motifs but also offer new opportunities for the digital preservation and documentation of local cultural heritage.

Actions (login required)

View Item