Figure 3
From: Tracking objects outside the line of sight using 2D intensity images
Object model and cost function used for tracking.
(a) photo of an object cut from white plywood and its representation as surface elements (surfels). Note that although we use a flat object for demonstration, our method is also capable of handling three-dimensional objects. (b) XY slice of the cost function for positional tracking, centered around the global minimum. With a perfect image formation model and in the absence of noise, the minimum (marked by cross) and the measured position of the object (marked by circle) should coincide at a function value of exactly f(p) = 0. Under real conditions, the reconstructed position deviated from the true one by a few centimeters and the minimum was a small positive value.
