DCT-domain translation for composing better pictures in the MDSQ-based multiple description video coding | IEEE Conference Publication | IEEE Xplore