Pengembangan Model Pembelajaran Mesin untuk Klasifikasi Citra Lukisan Menggunakan Self-Organizing Map dengan Library Minisom

  • Rangga Eka Nanda Institut Teknologi dan Bisnis Kalbis
  • Yulius Denny Prabowo Institut Teknologi dan Bisnis Kalbis
Keywords: bag of visual word, k-means, painting image, scale-invariant feature transform, self- organizing map


This research aims to develop a model to recognize painting types of figurative and non-figurative using self-organizing map (SOM) algorithm. This research used painting images from WikiArt which was formed into figurative and non-figurative types. Methods used in this research implements bag of visual words (BoVW) model to represent image features, SOM algorithm as a classifier, and incremental model as a software development method. Features of an image based on BoVW model formed using scale-invariant feature transform (SIFT) and K-means methods. The BoVW feature representation then classified using SOM which uses rectangular topology and gaussian neighborhood function. The result of this research is an application to recognize painting images with 83.3% accuracy.