Reading between the lines: Object localization using implicit cues from image tags | IEEE Conference Publication | IEEE Xplore