A computational model of selecting visual attention based on bottom-up and top-down feature combination | IEEE Conference Publication | IEEE Xplore