Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2021
- Goals
- Record
- 4af81df7-c5a6-4555-97cc-7e3fc0c75f09
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted an activity to develop a machine learning model for classifying batik authenticity using a Convolutional Neural Network algorithm with a transfer learning method. The activity involved Putra, Farrel Athaillah; Jamil, Dwi Anggun Cahyati; Prabandanu, Briliantino Abhista; Faruq, Suhaili; Pradana, Firsta Adi; Alya, Riqqah Fadiyah; Santoso, Heru Agus; Al Zami, Farrikh; and Saputra, Filmada Ocky as participants in the activity.
The activity addressed problems in the batik industry concerning the labelling of traditional and modern batik products. The source describes the prevalence of fraud involving printed batik sold at a price equivalent to written batik, which is much more expensive, as well as public ignorance of the aesthetic value and authenticity of written batik. These conditions can disrupt the traditional batik industry in Indonesia. Batik is also described as one of Indonesia's cultural heritages recognized by UNESCO as an Intangible Cultural Heritage.
The classification process consisted of dataset collection, data preprocessing, development of convolutional neural network models with transfer learning, and model compilation and training. The trained machine learning model produced an accuracy of 96.91%. The activity aimed to make it easier for people to distinguish between written and printed batik, minimize batik price fraud, and increase consumer confidence in batik transactions by ensuring the originality of batik products. Through this activity, UDINUS relates machine learning research to the preservation and responsible recognition of batik authenticity.