Clustering of Fine Art-Images as Digital Learning Content using Data Mining-Image Analysis Techniques

Abstract

Digital learning was currently packaged exclusively by adding images to attract the students' interest. The use of images as learning content is rarely well-presented as there was no attention given to how such images were produced. About this, Fine Art-Drawing Images are seen as a breakthrough; they have many types, and to use them as learning content, they must be clustered. Thus, this study aimed to determine and predict fine art-drawing images based on the characteristics of drawing techniques. The study used thirty images as input data in which each category (i.e., charcoal, chalk, pastel, pencil, pen and ink, and book illustration) consisted of five sample images. This study used the image analysis technique to process data with software orange data mining. This study revealed that there was a difference between the proportion of actual and proportion of predicted. Actual data were grouped by image technique only, while the prediction result considered image technique, texture, and image coloring. The similarity of drawing technique gives the result of predictive data, which is different from actual data. In a nutshell, this image analysis technique can be used to determine the cluster and prediction of images that have different characteristics and attributes.