Machine Learning Based Interactive Drawing Platform

Kural O. E., Kılıç E.

25th Signal Processing and Communications Applications Conference (SIU), Antalya, Turkey, 15 - 18 May 2017 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu.2017.7960668
  • City: Antalya
  • Country: Turkey
  • Keywords: Hand drawing shape recognition, Support vector machine, Mobile learning, (Modified) Hausdorff distance
  • Ondokuz Mayıs University Affiliated: Yes

Abstract

In this study, supporting mobile application based on machine learning in which the children can do drawing studies on basic geometric shapes is developed. Some feedback is provided to support the drawing of the children by creating a drawing platform in application. 5-class dataset consisting of straight line, triangle, rectangle, circle and ellipse shapes was created by choosing 6-8 years old children as target group for the study. Drawing data that classified with support vector machines were graded with Hausdorff distance criteria. Feedback messages supporting the drawing phases were generated according to the obtained grades.