Towards Scene Understanding: Unsupervised Monocular Depth Estimation With Semantic-Aware Representation | IEEE Conference Publication | IEEE Xplore