M12.7 CONNECT THE MECHANISM
Infer structure that a flat image does not directly reveal
Portrait mode blurs the background, so your phone must know how far away every pixel is. Work out depth from two cameras, motion from two frames, and meet NeRF and Gaussian splats.
LESSON OVERVIEW14 min lesson
Lesson overview
Portrait mode blurs the background, so your phone must know how far away every pixel is. Work out depth from two cameras, motion from two frames, and meet NeRF and Gaussian splats.
What you’ll explore
- Depth, optical flow, and 3D reconstruction estimate hidden geometry from visual evidence; camera models, multiple views, motion, and learned priors introduce essential assumptions.
GO TO THE SOURCE
Original explanations, connected to the research.
Computer Vision: Algorithms and Applications, 2nd edition, chapters on motion, stereo and 3D reconstruction (Szeliski, 2022)RAFT: Recurrent All-Pairs Field Transforms for Optical Flow (Teed & Deng, 2020)NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis (Mildenhall et al., 2020)3D Gaussian Splatting for Real-Time Radiance Field Rendering (Kerbl et al., 2023)Modern Robotics — authors’ textbook and resourcesSuggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.