A neighboring vector, which is a boundary portion between two overlapping objects, is extracted. To calculate luminance levels of the objects on both sides of the neighboring vector, a predetermined number of coordinate points (sample points) in the vicinity of the neighboring vector are extracted at least from the image side. A rendering process is performed on an area including all the extracted sample points to acquire color values at the sample points. The luminance level of the image is calculated based on the acquired color values, and the luminance levels of the objects on both sides of the neighboring vector are compared to each other to determine the position (direction) in which to generate a trap graphic.