Learning sparse representations by K-SVD for facial expression classification | IEEE Conference Publication | IEEE Xplore