Crafting Monocular Cues and Velocity Guidance for Self-Supervised Multi-Frame Depth Learning elfframeguididalocmonopthramsedselfsuperuidvelocity