Talk:Video tracking

Tracking Paradigms
I would like to add a "tracking paradigms" chapter:

Object tracking can be formulated in numerous ways. One line of research applies a pre-learned object model robustly to an image sequence. However, the sheer number of possible appearance changes make objects hard to render in advance. Adaptive methods are proposed in order to learn the object’s appearance on-the-fly. Alternatively, objects can be treated as outliers to a scene model which might be easier to learn, especially with a fixed camera. Furthermore, objects can be discovered using segmentation techniques.

The four tracking paradigms: pre-learned object models for tracking, adaptive models for tracking, scene models for tracking and segmentation models for tracking would each deserve their own section.

Source: Severin Stalder, PhD Thesis ETH Zurich 2012

Are there any objections on doing that?

Algorithms
I would like to change the "algorithms" chapter to "target representation" and "data association".

What do you think?

The introduction
The introduction is much clearer now.

I disagree : there are more than only 2 main methods for tracking. I add at least one

Also, "condensation" is more significant than "contour tracking" -- other visual features than contours can be used.

Page move
I do not agree this page to be merged with motion estimation. Although somehow related, they are different concepts and shall not be confused with each other. Video tracking aims to track an object in sequential video frames; motion estimation tells the difference between two adjacent frames, but this difference might not necessary be the movement of real objects in the video. Two techniques serve different purposes: video tracking is heavily associated with computer vision, helping a computer to recognize objects in received video frames; motion estimation has nothing to do with the object recognition, it is mainly used as a video compression technique (see MPEG2 ). It might be arguable that techniques implementing motion estimation are useful for video tracking, but they are still conceptually different. —Preceding unsigned comment added by Alienbat (talk • contribs) 13:08, 20 June 2009 (UTC)

Video tracking and motion analysis are totally different concepts...95.16.62.135 (talk) 11:25, 31 December 2009 (UTC)

Description of motion model for video compression is wrong
"For video compression, key frames are divided into macroblocks. The motion model is a disruption of a key frame, where each macroblock is translated by a motion vector given by the motion parameters."

This is misleading. Firstly, the motion model has nothing to do with key frames. Motion is defined for any pair of frames (A, B) where B is not intra-only. Secondly, motion is defined at the prediction block level (a subdivision of a macroblock). Thirdly, there can be two motion vectors per prediction block.

Laurent Birtz — Preceding unsigned comment added by 207.253.7.190 (talk) 19:42, 26 October 2015 (UTC)

A Commons file used on this page has been nominated for speedy deletion
The following Wikimedia Commons file used on this page has been nominated for speedy deletion: You can see the reason for deletion at the file description page linked above. —Community Tech bot (talk) 14:07, 18 November 2019 (UTC)
 * Signal triangulation.png