Implementation of Conditional WGAN-GP, ResNet50V2, and HDBSCAN for Generating and Recommending Traditional Lombok Songket Motifs
DOI:
https://doi.org/10.30871/jaic.v9i5.10894Keywords:
CWGAN-GP, ResNet50V2, HDBSCAN, Songket, Generation of MotifsAbstract
Songket is a traditional Indonesian woven textile with profound cultural and aesthetic value, particularly in Lombok, where artisans continue to preserve its distinctive motifs. However, the creation of new designs is still carried out manually, requiring considerable time and relying heavily on the artisans’ creativity. This study proposes an integrated system that combines Conditional Wasserstein Generative Adversarial Network with Gradient Penalty (CWGAN-GP), ResNet50V2, and HDBSCAN to automatically generate and recommend Lombok’s traditional songket motifs. The dataset consists of primary data collected directly from local artisans and secondary data from the BatikNitik public repository, thereby providing authentic yet diverse motif samples for training. CWGAN-GP is employed to synthesize motifs with stable and realistic structures across multiple epochs. Subsequently, ResNet50V2 is utilized for deep visual feature extraction, HDBSCAN for density-based clustering, and UMAP for two-dimensional visualization of motif distribution. The system successfully groups motifs into meaningful clusters, with the largest cluster containing consistent patterns of high aesthetic value. A recommendation mechanism is also developed to suggest up to five similar motifs from the original dataset within the same cluster, ensuring cultural relevance while fostering design innovation. Despite these promising outcomes, several limitations remain, such as the relatively small number of songket motif samples, variations in motif quality, and challenges during data collection including inconsistent lighting and non-uniform patterns. These factors affect both dataset consistency and generative performance. Nevertheless, this approach demonstrates the potential of artificial intelligence to support the preservation and innovation of cultural heritage by assisting artisans in creating and exploring new motifs more efficiently without losing their traditional identity.
Downloads
References
[1] H. Hambali, M. Mahayadi, and B. Imran, “Classification of Lombok Songket Cloth Image Using Convolution Neural Network Method (Cnn),” Pilar Nusa Mandiri, vol. 17, no. 85, pp. 149–156, 2021, doi: 10.33480/pilar.v17i2.2705.
[2] Z. Mutaqin and B. Imran, “Klasifikasi Kain Songket Khas Lombok Menggunakan CNN dengan Arsitektur Alexnet,” Explore, vol. 14, no. 2, pp. 108–112, 2024.
[3] M. Multazam and E. Y. Saniyah, “Development and Implementation of Woven Bamboo Handicraft Online Shop in Loyok Village, Lombok, Indonesia,” J. Techno Nusa Mandiri, vol. 17, no. 2, pp. 123–130, 2020, doi: 10.33480/techno.v17i2.1638.
[4] E. Wahyudi, B. Imran, S. Erniwati, M. N. Karim, I. Pemerintahan, and D. Negeri, “Fine-Tuning Resnet50v2 With Adamw And Adaptive Transfer Learning For Songket Classification In Lombok,” Pilar Nusa Mandiri, vol. 21, no. 1, pp. 82–91, 2025, doi: 10.33480/pilar.v21i1.6485.
[5] D. T. S. Kumar, S. Muthuvelammai, and N. Jayachandran, “AI in Textiles: A Review of Emerging Trends and Applications,” Int. J. Res. Appl. Sci. Eng. Technol., vol. 12, no. 11, 2024.
[6] S. Khalil, M. Taha, R. Ali, H. Mehmood, and H. Dilpazir, “Generation of Textile Patterns Through Generative Adversarial Networks,” in Conference: 1st International Conference on Software Engineering and Computing Disciplines(ICSECD), 2020, pp. 1–9.
[7] V. V. K. Reddy, S. Cherukuri, K. Vanaparla, and L. R. Avula, “Deep Feature Extraction for Fashionable Fabrics: Using ResNet50, MobileNet, and CNN,” Lect. Notes Networks Syst., vol. 1096 LNNS, no. March, pp. 417–429, 2025, doi: 10.1007/978-981-97-7178-3_36.
[8] B. Imran and M. M. Efendi, “The Implementation Of Extraction Feature Using Glcm And Back-Propagation Artificial Neural Network To Clasify Lombok Songket Woven Cloth,” J. Techno Nusa Mandiri, vol. 17, no. 2, pp. 131–136, 2020.
[9] R. A. Fayyaz, M. Maqbool, and M. H. Hanif, “Textile Design Generation Using GANs,” in 2020 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), 2020.
[10] M. Liu and B. Zhou, “Innovative Design of Chinese Traditional Textile Patterns Based on Conditional Generative Adversarial Network,” in Culture and Computing, 2022.
[11] G. N. A. H. Yar, M. Taha, Z. Afzal, F. Zafar, I. U. R. Shahid, and A. Noor-Ul-Hassan, “TexGAN: Textile Pattern Generation Using Deep Convolutional Generative Adversarial Network (DCGAN),” Proc. - 2023 IEEE Int. Conf. Emerg. Trends Eng. Sci. Technol. ICES T 2023, no. June, 2023, doi: 10.1109/ICEST56843.2023.10138848.
[12] S. A. Ahteck et al., “Generative AI for textile engineering: blending tradition and functionality through lace,” MIT Press (in Press., pp. 1–36, 2024.
[13] H. Simanjuntak, E. Panjaitan, S. Siregar, U. Manalu, S. Situmeang, and A. Barus, “Generating New Ulos Motif with Generative AI Method in Digital Tenun Nusantara (DiTenun) Platform,” Int. J. Adv. Comput. Sci. Appl., vol. 15, no. 7, pp. 1125–1134, 2024, doi: 10.14569/IJACSA.2024.01507109.
[14] A. E. Minarno, T. D. Antoko, and Y. Azhar, “Generation of Batik Patterns Using Generative Adversarial Network with Content Loss Weighting,” Int. J. Adv. Sci. Eng. Inf. Technol., vol. 13, no. 1, pp. 348–356, 2023, doi: 10.18517/ijaseit.13.1.16201.
[15] A. E. Minarno, I. Soesanti, and H. A. Nugroho, “Dataset of Batik Nitik Sarimbit 120,” Data Br., vol. 55, pp. 0–7, 2024, doi: 10.1016/j.dib.2024.110671.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Ardiyallah Akbar, Muh Nasirudin Karim, Bahtiar Imran
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
