Patch-level Representation Learning for Self-supervised Vision Transformers | IEEE Conference Publication | IEEE Xplore