Depth estimation produces a per-pixel distance map from a normal image, no depth sensor needed. It lets effects sit correctly in front of or behind objects, supports robotics and AR, and adds geometry to otherwise flat footage.
The book · Volume 6 of 7
AI for Video: A Complete Guide to Intelligence in Video Products: From Models and Inference to Computer Vision, Speech, Search, and Generative Video
Depth estimation adds geometry to flat footage, and errors at object edges are what users notice first in AR effects. This volume covers depth estimation in depth as part of its computer vision section: monocular depth models, segmentation and pose, model artifacts, and where to run inference on device or in the cloud. Written by Nikolay Sapunov, CEO at Fora Soft.
Need depth estimation from ordinary video for AR or scene analysis?
Describe the effect or measurement you need, and we will propose a depth estimation model and pipeline with accuracy and latency targets for your devices. 250+ video products shipped since 2005.
