Automatic matting using edge feature-based scribbles

Foreground and background extraction, or commonly called matting, is still widely used in computer vision applications. Most of the matting, constraints of the foreground are determined manually. However, it is more efficient if done without human intervention. We proposed a method to determine constraint, scribbles, automatically using edge feature on image matting. Edge features are used to find gradient vector from each RGB color channel to obtain automatic scribbles. Our result will compare with a manual process using Mean Absolute Error (MAE). All of the experiment in this research were used several images from some previous work. The experimental result shows our proposed method has only differed 0,0336 from a current state of the art research and it done automatically without human effort.