a)The first task is to build a Gaussian pyramid for an input image. Create a routine pyramid = MakePyramid(image, minsize) that creates a pyramid of images. It returns the original image as well as a sequence of images of reduced size, using a scale factor of 0.75 from one level to the next. The pyramid should stop when any further reduction in size will make a dimension of the image smaller than minsize. You can use the Matlab routineimresize(image, 0.75, 'bicubic') to reduce the image size at each step (bicubic interpolation avoids all aliasing problems).

b)In order to find instances of a template in a pyramid, write a function FindTemplate(pyramid, template, threshold). It finds all locations of the template in the pyramid at which the normalized cross correlation is above the given threshold. You can use Matlabs normxcorr2 function.

c)In order to draw detected locations, write a Matlab function ShowFaces. It generates and displays boxes using lines. You can use the Matlabs line function to draw lines (i.e., line([10 10], [20 100], 'color', 'r')).