Prediksi Motif Batik dengan Menggunakan Metode Gabor Filter Convolution Neural Network

Bili, Yudisman Ferdian
Tundo
Sutisna, Nandang
Putri, Atsilah Daini
Yuliantoro, Dita Tri
Nurmayanti, Laily

Jurnal JTIK (Jurnal Teknologi Informasi dan Komunikasi)
Vol 9 No 3 (2025): JULI-SEPTEMBER 2025

This research aims to develop a batik motif classification system by utilizing Convolutional Neural Network (CNN) and Gabor Filter, in order to increase accuracy in texture feature extraction. The batik dataset used goes through a preprocessing stage, which includes normalization and data augmentation. During training, the model was tested with 10,000 iterations, using the Adam optimizer and the Categorical Cross-Entropy loss function, and evaluated via a confusion matrix. Test results show accuracy reaching 87%, with a precision and recall value of 90% each, and an F1-score of 89%. This method has proven effective for classifying batik motifs and has the potential to be applied in the fields of education, textile industry and cultural preservation.