Tuesday, February 24, 2009

Video Annotations, Navigations, and Video-Still Composition

Videos use all kinds of technology today to make viewing more enjoyable. Annotations, Navigation, and Video-Still Composites are some of the most enjoyed luxuries. Dr. Dan Goldman proposes ways to improve all of them.

Video Annotations
Annotations are what make viewing easier and understandable. You see it every day: When watching a sporting telecaster draw motion, on youtube when a bubble pops up to describe what's going on, when you see thought bubbles on TV. John Madden made famous the idea of drawing routes receivers run, where a runningback goes, where a blocker takes his angle. Now, everyone uses it. You see it in tennis matches tracking the path of the ball, basketball, everything. On youtube, the video annotator allows the uploader of a video to add subtitles, speech bubbles, and other things.

Video Navigation
If you use VLC Media Player, Real Player, Windows Media Player, ANYTHING, you see a line that is start to finish. There is a timeline going through the movie with a marker letting the user know where they are in relation to the start and end of the video. They can drag the marker to a point on the timeline and the movie will jump to that point, as happens on any youtube video.

Video-Still Composition
We think of videos as compositions of many different frames which, when put together one after the other, form a seemless video. The problem comes when we take one still image; we can see tons of motion. Nowadays, cameras come with a burst function which take 2 to 10 images in a row, great technology.

Particle Grouping Approach
The approach of Dr. Goldman is particle grouping approach of Sand and Teller. It takes still images, and makes different particle groups based on what's in the picture, and is able to do many things with the particles, such as annotations of motions, graffiti, etc.



In this picture, as you can tell, it is divided into sections of particles which separate the two kids walking through the woods, and the woods themselves. Using some very complicated algorithms that I didn't really understand, we can do things such as annotate images with motions of the paths they take. For example, take the picture below:


If we make number 37 based on his particles, the video will automatically draw the path of motion he takes when we play it. There are other video annotations and navigations we can use such as graffiti where we would change his number from 37 to 3 and then can play the video, using the particles, as him being number 3, not 37. And we can navigate based on when he is at a certain point in the video. We can choose when number 37 is around the line of scrimmage and it will skip to that point in the movie, not when he is receiving the handoff 4 yards deep.

1 comment:

  1. Looks like a pretty interesting paper that you've read. Video manipulation techniques are pretty impressive and definitely benefits many programs and viewers.

    ReplyDelete