Meeting Marcel 15-June

I had two ways to go.

1. Compute the position and orientation of the camera for the time sequence. I have a demo already done for that. The problem is that in order to estimate the fundamental matrix, I need to assume depth, which is contrary to one of our main assumptions of planarity of the scene.

2. The second one is the one that I will go now for. So I will track the scene, using the homography as state vector. The temporal model would be brownian motion.

 

I have started writing the abstract and introduction.