LaPE: Layer-adaptive Position Embedding for Vision Transformers with Independent Layer Normalization | IEEE Conference Publication | IEEE Xplore