Tracking Practice: Putting An Image Onto A TV

Tracking is a vital part of my assessment, in particular the car colour correct part, the one where the image pans down. This is a blog post showing how some of the tracking material from the week 7 lectures can be used as practice for the final brief piece. The material involves tracking some moving footage of a television, and then using the cornerpin node to attack the corners of the second image to the tracking points.

Then read in the image that is to be put on the television

Using the cornerpin node, copy the information from the tracker so that the new image fits in properly with the television corners:

That should lock the image to the trackers, but for some reason it didn't, so bring in a transform node to scale the image to the size it needs to be.

In order to make the TV look more realistic, blur the edges of the screen.

Now the image should look like it was always there:

It looks pretty good apart from the top corner which for some reason jumps.

The same could also be done with a video:

It actually appears to work better with a video.

If this piece were to be done again, something that would need some attention is the problems with the cornerpin node, it should have instantly attached the corners of the image to the tracking points. Also the problem with the corner of the image jumping will need looking at. Finally the edge blur on the Michael Fassbenger interview can be edited slightly, it looks a bit grey.