The identification technology enables a user to easily cut out specific areas within an image displayed on a computer or a tablet. This technology allows users to, for example, cut out an object within an image by running a fingertip over the general area of the object, which is then identified.
In order to identify the borders of the selected object in the image, this identification technology then "labels" each pixel as either "red" or "blue," starting from the pixels that were traced by the user's finger (shown as red and blue lines in the figure). An algorithm attempts to continue labeling each adjacent pixel in the same color. When the two colors collide and "fight" over a pixel, specific conditions determine which color that pixel belongs to. Finally, all pixels are labeled and the borders of the object are identified (Fig. 1 Top). Multiple objects can also be cut out simultaneously using the same algorithm (Fig. 1 Bottom).