Self-Supervised Deep TripleNet for Video Object Segmentation | IEEE Journals & Magazine | IEEE Xplore