A DCT-domain JND model based on visual attention for image | IEEE Conference Publication | IEEE Xplore