MonoMAE: Enhancing Monocular 3D Detection through Depth-Aware Masked Autoencoders | Read Paper on Bytez