Self-supervised Audio-visual Co-segmentation | IEEE Conference Publication | IEEE Xplore