To identify the content of the images, the SURF (SpeedUp
Robust Features)[2] algorithm has been used. It extracts
a number of unique and repeatable key points between
different images of the same content and is also immune to
rotation and scaling (see figure 1). The neighborhood of
pixels of these key points is coded into vectors. The
comparison of these vectors of different images determines
the existence of a match between them. The images should
be compared one by one, so it is useful to compare the
lowest possible number of images. We therefore focus on
those places that are visible to the user to reduce the number