00:00:00
|
Lecture 21: Structure-from-Motion
|
00:01:29
|
Announcements
|
00:08:36
|
Announcements (2)
|
00:11:39
|
Course Outline
|
00:12:01
|
Recap: Estimating Optical Flow
|
00:12:50
|
Recap: Lucas-Kanade Optical Flow
|
00:14:31
|
Recap: Iterative Refinement
|
00:15:27
|
Recap: Coarse-to-fine Estimation
|
00:17:13
|
Topics of This Lecture
|
00:18:37
|
Structure from Motion
|
00:19:45
|
What Can We Use This For?
|
00:21:53
|
Structure from Motion Ambiguity
|
00:25:03
|
Reconstruction Ambiguity: Similarity
|
00:26:07
|
Reconstruction Ambiguity: Affine
|
00:26:42
|
Reconstruction Ambiguity: Projective
|
00:27:13
|
Projective Ambiguity
|
00:27:52
|
From Projective to Affine
|
00:28:31
|
From Affine to Similarity
|
00:29:28
|
Hierarchy of 3D Transformations
|
00:32:01
|
Topics of This Lecture
|
00:33:54
|
Structure from Motion
|
00:35:19
|
Orthographic Projection
|
00:36:39
|
Affine Cameras
|
00:39:23
|
Affine Structure from Motion
|
00:51:07
|
Factorizing the Measurement Matrix
|
00:54:32
|
Affine Ambiguity
|
00:56:09
|
Estimating the Euclidean Upgrade
|
01:01:41
|
Algorithm Summary
|
01:02:14
|
Reconstruction Results
|
01:03:04
|
Dealing with Missing Data
|
01:07:05
|
Comments: Affine SfM
|
01:08:47
|
Topics of This Lecture
|
01:08:51
|
Projective Structure from Motion
|
01:11:02
|
Projective SfM: Two-Camera Case
|
01:19:23
|
Projective Factorization
|
01:20:54
|
Sequential Structure from Motion
|
01:22:51
|
Bundle Adjustment
|
01:25:23
|
Projective Ambiguity
|
01:26:14
|
Self-Calibration
|
01:27:07
|
Practical Considerations (1)
|
01:27:49
|
Practical Considerations (2)
|
01:28:50
|
General Guidelines
|
01:29:30
|
Commercial Software Packages
|
01:29:42
|
Applications: Matchmoving
|
01:30:01
|
Applications: Large-Scale SfM from Flickr
|
01:30:33
|
References and Further Reading
|