ModelDense Prediction Transformer
DPT
A Transformer-based monocular depth estimation model developed by Meta, used to infer depth information from a single RGB image. It is widely applied to depth map extraction tasks in AI video generation pipelines.
All Facts (2)
Source Articles
A New Breakthrough in Parallel Thinking: How DPT Makes LLM Reasoning Paths More DiverseOct 22Monocular Depth Estimation in Practice: How a Self-Trained Model on 750K Images Enables 3D Reconstruction with a SmartphoneOct 6Depth Map and OpenPose Extraction Workflow: A Practical Guide to Controllable AI Video GenerationJul 12