EE 409 – Mini Project 4 Due: 03.01.2017, 22:00 (Wednesday) Notice: A severe penalty will be applied in case of cheating 1. (Finding a suitable image file) For this mini project, you will again use a photo of yourself as you did in the previous mini projects. Important: The resolution of the image should be higher than 0.1 megapixel and lower than 1 megapixel. Don’t forget to put the Matlab codes and the obtained images in each of the following questions to your report. Hint: You should obtain figures similar to the ones given next to the questions. 2. (Distorting image by blurring in spatial domain) Blur your grayscale photo using an "n x n" averaging filter in spatial domain. If your image has "c" number of columns then n=2*round(c/100)+ 1 . (Ex: If your image has 437 columns, then n=2*round(437/100) + 1 = 9; So apply 9x9 averaging filter to your image) 3. (Distorting image by blurring in frequency domain) Now take the fft of your grayscale image. Then blur your photo(using its fft) with an "n x n" averaging filter in frequency domain. (You should use the fft of the averaging mask). Then take the inverse fft and display the blurred image.
4. (Reconstructing blurred image by inverse filtering) Deblur your blurred image in frequency domain. (Note: You should apply edgetaper() function to your blurred image as given in examples to smooth the image border). Compare the deblurred image with the blurred image and comment on the result.
5. (Adding noise to the blurred image) Add very small amount of uniform random noise in the range [-0.02, +0.02] to your blurred image.
6. (Reconstructing noisy blurred image by inverse filtering) Deblur your noisy blurred image in frequency domain. (Note: You should apply edgetaper() function to your noisy blurred image as given in examples to smooth the image border). Compare the noisy deblurred image with the deblurring result in question 4. Comment on the result and its reason. 7. (Image compression using fft) You will compress your image by deleting some of its highest frequency fft coefficients. Preserve only half the size of rows and columns of coefficients in the center of the image fft. (Don't forget to use fftshift and ifftshift). Compare the quality of the compressed image with the original image and comment on the result. How much space does your compressed image fft take if your original image fft takes 100?