Homework Assignment for ECE 456/556: Introduction to Pattern Recognition - Moments, Assignments of Electrical and Electronics Engineering

Download Homework Assignment for ECE 456/556: Introduction to Pattern Recognition - Moments and more Assignments Electrical and Electronics Engineering in PDF only on Docsity! ECE 456/556 – Introduction to Pattern Recognition Spring 2009 Homework #2 due Thursday 29 January Moments 1. Using MATLAB there is a function available to you to investigate moments. To access this, (1) start Matlab (2) add the path by typing addpath(‘m:\Labs\Ee\Pattern_rec\Intro_PR_functions’); (3) start the GUI by typing moments after the ‘>>’ prompt. Investigate first how each moment relates to the physical images. Change the image bitmap and see how the shape of the character changes each of the 10 moments. Qualitatively how do each of the moments change if a single pixel is filled in/deleted from (a) the top row, (b) the bottom row (c) the right side (d) the left side (e) the middle? 2. Try to see that different images/characters give different sets of moments and perhaps get a feel for what shapes have what types and sizes of moment values. Which moments are changed for a tall thin image vs. a short thin one (caution do not make the width = 1, or the moments will be 0 which is indeterminant)? a tall fat vs. a tall thin one? one with the mass near the center vs. far from the center (rows of 0110 vs. 1001)? one with most mass on the right vs. on the left (rows of 11001 vs. 10011)?. 3. The formulae for calculating moments is given in R. C. Gonzalez and R. E. Woods, Digital Image Processing, Prentice Hall, second edition, page 672-675. Read this description. Calculate moments m00 m01 m10 μ02 μ11 μ20 μ03 μ12 μ21 μ30 by hand of the following 3x2 pattern: 1 1 1 1 0 1 4. Select a programming language and platform which you will use for your programming projects in this course. Download the data file train.dat from the course website. This file contains the 8 central moments we discussed in class for 10 characters from each of 10 character classes. Write code to read in this data. 5. The Root Mean Square (RMS) is a common statistic about features. Its formula is N data RMS N i i∑ == 1 2 . It is literally the square root of the mean (average) of the square of the data. We will use this to normalize the data used in the projects. Write code and calculate the RMS value for each moment over all classes (8 values). Submit your code (well commented) and these 8 calculated values.