Global Motion in Computer Vision - Lecture 9

Explore global motion in computer vision through this comprehensive lecture. Delve into motion models, basic components, and coarse-to-fine global flow estimation techniques. Learn about image warping, video mosaics, and sprite creation. Examine projective flow, both weighted and unweighted, as well as bilinear and pseudo-perspective approaches. Discover how to determine projective transformation using point correspondences. Gain insights from Dr. Mubarak Shah's expertise in this hour-long presentation from the University of Central Florida's Computer Vision Video Lectures series.

Intro
Motion Models
Global Motion
Contents
Basic Components
Coarse-to-fine global flow estimation
Estimation of Global Flow
Image Warping (Bergan et al)
Video Mosaic
Sprite
Projective Flow (weighted)
Projective Flow (unweighted)
Bilinear and Pseudo-Perspective
Algorithm-1
Determining Projective transformation using point correspondences