Learning to Prompt CLIP for Monocular Depth Estimation: Exploring the Limits of Human Language | IEEE Conference Publication | IEEE Xplore