Lecture-30

Lecture-30 - 1 Classification of Video Shots Using 3D...

Info iconThis preview shows pages 1–5. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 Classification of Video Shots Using 3D Camera Motion Using 3D Camera Motion August 2005 VACE Review Task Definition ! Goal: Given the video sequence categorize video ! Goal: Given the video sequence categorize video shots into the following three categories: 1. Pan/Track: horizontal motion 2. Tilt/Boom: vertical motion 3. Zoom: either zoom-in or zoom-out ! Testing Dataset: TRECVID 2005 dataset August 2005 VACE review 1. Broadcast videos 2. MPEG-1 3. 140 videos (30mins to 1 hour each), 200~400 shots per video. 2 Motion Models (long videos) Pan (right) Tilt (up) Zoom (out) August 2005 VACE review Track (right) Boom (up) Static Motion Models (short videos) Pan (right) Tilt (up) Zoom (out) August 2005 VACE review Track (right) Boom (up) Stati c 3 Motion Models (a). Pan (Left) (c) Tilt (Down) (b). Track (Left) (d) Boom (Down) August 2005 VACE review (c). Tilt (Down) (e). Zoom (Out) . (f). Static Common Approach ! Based on the analysis of the optical flow field; (2D motion) ! Computationally Expensive; August 2005 VACE review ! Cannot distinguish pan/tilt from track/boom. 4 Proposed Approach ! Uses 3D geometric analysis ! Homography: pan, tilt and zoom; ! Fundamental matrix: track and boom....
View Full Document

Page1 / 9

Lecture-30 - 1 Classification of Video Shots Using 3D...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online