Computer-Vision-Warmup

Basic Computer Vision operations on an image file using MATLAB

1. [1 point] Load the image peppers.bmp into a variable A. Display the loaded image A on figure 1 with the message “RGB Original Image” as the figure title. {Think: What is the data type of A? What is the size of A?} Matlab hints: imread, figure, imshow, title, disp, pause

2. [5 points] Convert image A into a grayscale image and store it as B. Transpose image B as TB. Horizontally flip image B as HB so that the upper half of B becomes the lower half of HB and the lower half of B becomes the upper half of HB. Flip columns of image B in the left/right direction as FB.

Display images B, TB, HB, and FB on figure 2 with B located at the upper left, TB located at the upper right, HB located at the lower left, and FB located at the lower right. Label each image with its corresponding matrix name (e.g., B, TB, HB, and FB). Matlab hints: rgb2gray, transpose (or ’), subplot, fliplr, flipud, flipdim

3. [10 points] Save the maximum, minimum, mean, and median intensity values of B in appropriate variables by calling appropriate Matlab built-in functions, respectively. Write a Matlab function FindInfo to calculate the maximum, minimum, mean, and median intensity value of a grayscale input image. Inside FindInfo function, you are not allowed to call max, min, mean, and median functions. In other words, you have to write your own solutions to get different statistics computed. The prototype of FindInfo function is as follows: function [maxVale, minValue, meanValue, medianValue] = FindInfo(oriIm); where oriIm is the original grayscale image. Call FindInfo function in your main script to compute maximum, minimum, mean, and medianvalues of B and save the computed intensity values in appropriate variables, respectively. intensity Compare your computed four statistics with the four statistics obtained from the Matlab built-in functions using a series of “if” or “if ... else” statements and print the comparison results on the Matlab console. Note: The results computed from your function should be the same as the results computed from calling Matlab built-in functions. If they are not the same, either your implementation is wrong or you called the wrong Matlab built-in functions or called the Matlab built-in functions in a wrong way. Please fix the problems before submitting your assignment.

4. [4 points] Normalize image B to C, whose data type is double and whose values fall in the range of [0, 1] (i.e., the maximum intensity value of the image should be normalized to 1). Display image C on figure 3 with the message “Normalized Grayscale Image” as the figure title. (Note: Image C should appear the same as the image B.) Raise each pixel in the second quarter columns of image C to the power of 1.25 and raise each pixel in the fourth quarter columns of image C to the power of 0.75. Keep the first and the third quarter columns of image C unchanged. Store the result as an image (matrix) D. Display images D on figure 4 with the message “Processed Grayscale Image” as the figure title. You are not allowed to use loops to accomplish the task. Save image D in jpeg format to a file called “X_D.jpg” where X should be youra standard image viewing program to verify that the image is saved properly. first name.Open it using Matlab Hint:5. [5 points] double, /, ./, ^, .^, imwrite, display, disp, :

5. [5 points] Perform binary thresholding on the original normalized grayscale image C. A threshold 0.3 is chosen and all values in C greater than the threshold are set to 1, otherwise set to 0. Find two efficient solutions to obtain the thresholded binary image and save it in bw1 and bw2. Both solutions should not use any loop structure, should not call Matlab built-in function im2bw or imbinarize, and should be distinct in nature. Use the Matlab built-in function im2bw or imbinarize to do the same task and save its thresholded binary image in bw3. Compare your results bw1 and bw2 with the Matlab’s result bw3. If they are equal, display the message “Both of my methods worked”; otherwise, display the appropriate message such as “My method 1 worked but not my method 2”, or “My method 2 worked but not my method 1”, or “Both of my two methods did not work”. Of course, the first message should be displayed if your solutions are correct. Make sure that you do consider four aforementioned conditions in your coding. Display bw1, bw2, and bw3 side-by-side on figure 5 and label“my second method”, and “Matlab method”, respectively. the three images with “my first method”, Matlab Hint: find, >=, zeros, ones, &, &&

6. [5 points] Write a Matlab function BlurImage to replace each of 4 pixels in a non-overlapping 2×2 block of any input image with the average intensity value of these 4 pixels. For simplicity of coding, you can assume that the number of rows and the number of columns of an input image are divisible by 4. The prototype of BlurImage function is as follows: function [blurredIm] = BlurImage(oriIm); where oriIm is the original color image or grayscale image and blurredIm is the blurred color or grayscale image. Note: blurredIm and oriIm should have the same data type and the same size. The implementation of BlurImage function should contain one section of code to handle both color and grayscale images (similar to the polymorphism concept). Call BlurImage function in your main script to blur image A, and save the blurred image to variable BA. Call BlurImage function in your main script to blur image B, and save the blurred image to variable BB. Display images A, B, BA and BB in the raster-scan order (left to right and top to bottom) with the appropriate title on figure 6.

7. Close all figures and clear all variables. Matlab Hint: close, clear