James Du (jsd46)
Peter Greczner (pag42)
Compile time was the main bottleneck to our progress. We should have left our project broken up in small parts even after testing. Each part was developed on its own in a separate module; however, after a part's integration into the main project, we deleted or lost the individual modules and were unable to easily backtrack. Initially we were very satisfied with our green screen threshold, so we hard coded it into the main module.
If we'd had more time we would've worked more on refining keystroke recognition and switching to an Ethernet port so we could possibly stream in video to replace the green screen instead of just a still image.
The green detection seems to be almost as good as those used by weathermen. If the image quality were better, the reverse green screen would be up to standards since green detection is perfect. If we were tuning this effect for photography, we would lower the frame rate and increase image quality for the input from the camera.
We were satisfied with our green screen results. It is obviously not up to production standards, but it is good enough for recreational special effects.
We were very satisfied with the results of the serial communication. Although initially we struggled very much with getting accurate delivery of information we remained persistent in researching and find better ways to implement the receiving module. Eventually through research and much testing we implemented what we believe to be a very robust and reliable method of sending data following RS-232 protocol standards. The fact that we implemented hardware error detection and timeout errors was a great success for the project.