Self-Supervised Image-to-Point Distillation via Semantically Tolerant Contrastive Loss | IEEE Conference Publication | IEEE Xplore