Included is my code in the Code folder. Segment.cpp is the code used to segment the files, while magic.cpp was a simple script that output files in the pgm, grayscale format (equivalent to typing "convert " command line).

In order to compile segment.cpp and magic.cpp, type magick++-config. That will output an example of what should make magick++ files compile on your computer. Substituting segment and segment.cpp will allow for the usage in the project filed, namely being able to type the command "segment" followed by the image name.

/ The algorithm / The algorithm itself is simple. It inputs the image, moves it to the YUV colorspace, and then gets its width and height. Using those values it checks each pixel to see if it has a segment assigned to it. If it doesn't it increments the segments and assigns it to the pixel. If it does it just continues. Next the pixel is compared to four of its neighbors (this was an optimization since comparing four pixels assures that every possibility is covered so testing all neighbors is doing extra work).

When comparing if the neighbors have the same color value and have not been assigned yet, they are assigned the pixels segment. If they are the same color and have been assigned, that value is added to a chain stored in an array which means the two different segment numbers are actually members of the same segment (so if segment 2 is the same as 4, it would look like this- segment[2]=4). If the numbers are different the pixel is left alone.

Afterwards when counting segments the variable joins keeps track of how many distinct segments there are. The counter goes through all of the segments found one at a time. For each if they are part of a chain, the chain is followed and marked so that when that segment is checked we know not to add it as a separate segment (join in this case). Also chains pointing at the same end value which originally increment joins then decrement the value since they are joined by the last number in the chain.

Finally the number of actual segments (joins) is output on the image and it is written.