From bottom-Up visual attention to robot action learning | IEEE Conference Publication | IEEE Xplore