Load and visualize the image septagon.tif. (you can use any .tif file you want)

(a) Add Gaussian noise with mean 0 and variance 10 to this image, and visualize the noisy image and its histogram using the following code: septagon = double(imread(septagon.tif)); noise = 10*randn(size(septagon)); septagonNoisy10 = septagon + noise; figure;imshow(septagonNoisy10/max(septagonNoisy10(:))) figure;histogram(septagonNoisy10) What do you observe?

(b) Choose an appropriate threshold T and segment the image using the following code: septagonSegmented = imbinarize(septagonNoisy10,T); figure;imshow(septagonSegmented) Have you managed to successfully segment the image?

(c) Repeat part (b) when the noise variance is 20. Can you successfully segment the image in this case? Is there any T value that will result in an acceptable segmentation performance?

(d) Denoise the image you have generated in part (c) before segmentation, using the following code: h = ones(3)/9; denoisedSeptagon = filter2(h,septagonNoisy20); figure;imshow(denoisedSeptagon/max(denoisedSeptagon(:))) figure;histogram(denoisedSeptagon) septagonSegmented = imbinarize(denoisedSeptagon,T); Have denoising resulted in better segmentation performance?