Final Documentation

In the final production of my project, I have successfully implemented all three of the ofxOpenCv images classes (ofxCvColorImage, ofxCvGrayscaleImage and ofxCvFloatImage). These different ofxCv classes each functions differently through the same input of the camera. The output is different because of the restrictions I have set for each. Only when a certain degree of pixels will be shown once the subject moves a certain way. One of the most interesting thing I find doing this project is that how one webcam can display so many different things at once on one screen. Another thing that is fascinating is to observe the amplified pixels and also the buffer image. Sometimes if the subject doesn’t move at all in front of the webcam, then certain frames are completely dark. However, once the subject starts moving then these frames are back alive again. The way I have presented “pop art” on my live canvas is somewhat inspired by Andy Warhol’s Marilyn Monroe. I wanted to create something that is offset the grid and somewhat disturbing. Hence the fact that I used random color flashing in the background making this installation more strange.

An installation scenarios for my project would be projecting the display on a large scale. I would like users to experiment the different frames and see that one input can have multiple outputs. When users are moving so fast or so slow these frames can really show significant difference. What is fun about the last frame, which is the largest one of them al is that pixels are constantly moving around the subject. It doesn’t matter if they subject is moving or not, these pixels will float like small particles. It is fun to look at them and think what is the internal logic behind this. When the subject is not moving, it does not detect the pixels so then two of the frames will become dark. It is almost as if the subject is in control what is happening on the screen. One thing I would like further do with this project is to add in the morphing or other functions in the ofxOpenCv addon. It would be really great to see more happenings with this one input and many output system.

Final Prototype and ofxAddon Research

Since my concept is mainly focusing on the usage of webcam and experimenting the possibilities displaying multiple frames of different display all at once, I have found ofxOpenCv useful for my project. Open Computer Vision (OpenCV) library contains hundreds of classic and newest algorithms for image processing and analysis. In the ofxOpenCv addon there are collection of classes and these classes’ names begin with ofxCv. In my project I will be using ofxCvColorImage, ofxCvGrayscaleImage and ofxCvFloatImage for the different frames I want to explore in my display.

The ofxOpenCv addon supports my conceptual experimentation because I want to have different types of display and degrees of pixels shown on my webcam display. Not only do I get to manipulate how I want the gray scale to be but also having the variety and differences between each frame. I want to make four frames to mimic pop art style, however, in my project I would want to create one frame that is off centered and larger in scale than other frames. I want to go out the boundary of symmetry and also the conventional pop art canvas. I want to show these frames with offset coordinate and let users experience the oddness.

The ofxOpenCv addon works because I am calling the ofxCvColorImage which represents the three-channel color images with color components. Basically, this first frame wii just be what the camera reads. My next frame will use ofxCvGrayscaleImage class which will represent the grayscale with pixel values. These images will have internal processing such as thresholding and contour finding and also has the ability to work with binary images which contains only two pixel values. For the gray scale image, it will display the absolute difference of the frames. The third and fourth frame, it will use ofxCvFloatImage which can amplified difference image and the other one showing buffer image.

From playing with these ofxOpenCv addons, I have learned that these image classes can have varieties of display which can be further explored. My ultimate goal is to have these different displays all sync under one screen and let users experience and see for the difference in each of these frames.

Work Cited

http://opencv.org/