functional augmented reality

Camera pose estimation is very useful to create amusing toy applications, allowing the insertion of virtual objects in the scene. Let's show a simple example. First we take a sheet of paper with known aspect ratio (A4), which serves as our 3D reference frame:

When the rectangle is properly detected we insert the desired virtual objects as seen by the estimated camera:

Interestingly, the objects are actually "alive" in their small virtual world. If we lift the book the ball moves down and bounces against the walls:

There are many variations, games, etc, based on this idea. This kind of application is a good practical exercise for our computer vision students.

Here is video showing how it works under extremely adverse conditions: a cheap webcam, poor light, etc.

strange loop

In the first post I would like to show something related to the perceptual strange loops described by Hofstadter.

We have a simple application for contour recognition based on Fourier properties. The screenshot shows the raw contours in blue and the corresponding affine-normalized versions in yellow. We have clicked on one "A" and then all the similar shapes are shown in red:


The interesting thing is that the system also recognizes an "A" on its own "output" (the bottom-right window). In some sense, it sees that it is seeing an "A".

I'd like to work on this, trying to create some kind of feedback loop which reinforces itself and stabilizes the perceived entities.