Dataset Endek GAN
Description
The Endek dataset consists of 156 unique motifs of Endek, a traditional woven fabric from Bali, Indonesia. Each motif represents distinct visual patterns and color compositions that reflect the cultural and artistic heritage of Balinese textile design. The dataset has been curated to capture the diversity of Endek motifs, including geometric, floral, and symbolic patterns that are commonly used in ceremonial and contemporary contexts. This dataset is specifically prepared for research in Generative Adversarial Networks (GANs) and other machine learning applications, enabling tasks such as motif generation, style transfer, pattern classification, and cultural heritage preservation. The availability of 156 motifs provides sufficient variation to train and evaluate deep learning models, while also serving as a valuable resource for digital textile innovation and computational art studies. This research and dataset development were funded and supported by the Ministry of Higher Education, Science, and Technology (Kemdiktisaintek) of the Republic of Indonesia through the Directorate of Research and Community Service, Directorate General of Research and Development.
Files
Steps to reproduce
The dataset was collected by photographing Endek fabrics using a digital camera in various locations across Bali, Indonesia. The images were taken directly from woven fabrics available in traditional markets, weaving centers, and local communities to ensure authenticity and diversity of motifs. By capturing photographs from different regions in Bali, the dataset reflects a wide representation of Endek’s cultural and artistic variations. This approach preserves the originality of the fabric patterns while providing high-quality data suitable for machine learning and generative model research.