A15

Pattern Recognition

In this activity, we were asked to perform classification of different images through pattern recognition. Half of these images will serve as training sets, whereas the other half will serve as test sets. These training set are used to distinguish one class from another.

A14

Color Image Segmentation

In image segmentation, a region of interest (ROI) is chosen out of the rest of the image so that further processing can be done on it. Selection rules are based on unique features of the ROI.

There are 2 kinds of image segmentation. Paraemtric

and Non-parametric

Example 1. Original image

Our ROI

Image segmentation using Parametric segmentation
Image segmentation using Non-Parametric segmentationExample 2.

ROI

Image segmentation using Parametric segmentation
Image segmentation using Non-Parametric segmentation

Example 3.
ROI
Image segmentation using Parametric segmentation
Image segmentation using Non-Parametric segmentation
Based on our three examples, the Non-parametric segmentation gives a more accurate result compared to the parametric segmentation.

Score: 8/10. late posting

a13

Image Compression

Image to be compressed and then we convert it to grayscale (89.9KB)

Then we cut the image into 10 x 10 blocks and reshape this block into a 1-D array. We'll have an n x p matrix where n is the number of 10 x 10 blocks and p is the number of elements per block (100 in this case). Next we use the PCA method in scilab.

The figure below shows the correlation circle
Eigenvalues
The new correlation circle
New eigenvalues
Image using 3 out of 100 eigenvectors (19.6KB)
Image using 50 out of 100 eigenvectors (20.1KB)
Image using 100 out of 100 eigenvectors (22.6KB)

score: 8/10. i was able to do the activity, unfortunately, I posted late

A12

Preprocessing text

Original image to be processed

Our region of interest. It is tilted. Unfortunately, I could not get the mogrify code to work, so all the images are titled.
FT of ROI
FT of the image correlated with the filter
Black and white of the image at threshold value of 0.2
inverse of the black and white image
after dilation and erosion
this should be an image of the file that is not tilted. But since mogrify somehow doesn't work, this will have to do.
Black and white at threshold of 0,4
inverse of the black and white image
Description image used

score: 7/10. late posting yet again

A11

Playing notes by image processing

Mary had a little lamb music sheet (edited)

first line of the song

second line of the song
Halfnote
quarternote
correlation of line 1 with quarternote

converted to black and white witha threshold value of 0.8

line 1 and its FT

quarternote and its FT
correlation results
histogram of note spacing
end results: https://dl-web.dropbox.com/get/mary1.wav?w=d4c53403

score: 8/10

A10

Binary Operations

Original image cut into 12 256x256 subimages

average area is 530

New images after dilation and erosion

score: 6/10. I wasn't able to do the second part.