Progress Report 1

We have managed to make some real headway on our project this week!

Not only did we manage to learn a lot of fun little functions that we can use in our project, but we tried them out with a sample picture, and it actually works (sort of).

This is the picture that we chose to deal with: (It's an old-fashioned movie camera!)
>> [X,MAP]=gifread('movie.gif');
>> colormap(MAP);
>> image(X)

Then, we extracted out one of the movie reels from that picture.
>> reel=X(10:22,18:31);
>> colormap(MAP)
>> image(reel)

To do the pattern matching, we employ a 2-D convolution. Matlab has some useful built-in functions:
>> cnv_test=conv2(X,reel,'same');
>> colormap(gray(max(max(cnv_test))))
>> image(cnv_test)

The image is located at the bright white portions. I had to shrink the image to put it into there, and so it looks bad.

Then we went on to try to find a different part of the same image.
>> box=X(24:35,20:42);
>> figure(2)
>> image(box)

And the result of that one was:
>> cnv_test2=conv2(X,box,'same');
>> colormap(gray(max(max(cnv_test2))))
>> image(cnv_test2)

This is not a good representation, since Owlnet is not allowing me to have Netscape, XV and Matlab open simultaneously. Besides, the answer is incorrect since it is picking up the reels rather than the box.

This will be the focus of our work next week.