In order to distinguish our target from the background, we have made use of a color histogram for describing our targets. The histogram is made from the target’s image which has reduced resolution for computational saving purposes, from 256 available values per color channel to only 8 values per color channel.  An example of this is figure 1 below, where the image has been reduced to its pixels for analysis.  The descriptive model is based on the Red-Green-Blue(RGB) color distribution within the target, which is quantized to 512 available colors and recorded as a histogram.  A 3D representation of our color histogram can be seen in figure 2 for the previous example image. All of the colors that exist within the image appear in their RGB coordinates and the spheres radius of each color is representative of the number of occurences. This can be used to visualize how the algorithm uses the frequency of each color in the region of interest to describe the target and thus track it by matching this initial description.

Figure 1: Image Blur Figure 2: 3D Color Histrogram