MirroID is an installation piece, which aims to interact with those who look into the ‘mirror’. Inspired by the work of Francis Galton it uses real time face recognition and the process of composite photography to generate a snapshot of identity within a public space.
The project itself is an experiment, which aims to investigate the relationship between the camera, reflection and identity. The experience of looking into the ‘mirror’ aims to raise social and cultural questions surrounding the use of facial recognition technologies within public spaces.
In case you are wondering - The left hand side of the screen is the live web cam feed and the right is the composite snapshot.
After a review of my project today, it has been suggested i look into face recognition using AS3. I have found some cool posts but it looks quite complex for a beginner. However i am willing to have a go. I have found a source that uses real time face recognition to capture a snapshot of just the face. If i could implement this into my project it would be much better. Here is a demonstration video.
This got me thinking about how these pictures may be composited on top of one another. I took a few images from the FlickrBooth set on Flickr and then played with them in Photoshop using the hard light blend mode to get an idea of what it may look like. I cropped each photo similar to how the recognition works and made then all the same size.
I have been playing around with BlendMode and the filters in Flash to find out which one gives me the best result. I aim to generate composite snapshots of people across time so the layered bitmaps need to overlay one another, similar to using layers in Photoshop. Below there are some screen grabs from my project. BlendMode.HARDLIGHT has currently proven to be the best filter.
The last couple of days have been tough, with lots of help from Kseniya I have managed to learn some AS3. My project now:
1. Pulls in a live web cam feed to the stage.
2. Detects motion.
3. If motion is detected it takes a snapshot.
4. It then applies BlendMode.HARDLIGHT to the captured bitmap. This is so the snapshots can be layered to create a time-based image.
5. An array determines how many are layered at one time. I have currently set it to three.
I now have different variables - tolerance, filters and number of bitmaps - to experiment with. The foundations seem to be in place but it’s not quite complete.
Similar to the work of Francis Galton, Nancy Burson generates mixed portraits on the computer.
Below is “Big Brother (Stalin, Mussolini, Mao, Hitler, Khomeini, 1983″.
Nancy Burson’s pictures are products of that mentality. And they show it: they show what sort of world will result from that mentality: a world of chimaeras.
The heart desires something: for example, an androgynous human being, or an accomplished Big Brother, or a perfect female beauty. Such a desire may be called a ‹program.› In order to realize such a program in the form of an image, various available photographs should be chosen. [...in this case] the photos of Hitler, Mussolini, Stalin, Mao and Khomeini. [...] Each of these photos is to be calculated: dissected into very small squares called ‹pixels.› These calculations are to be fed into a computer, which is then instructed to compute the desired image from the pixels according to the program called ‹desire.› The result will be Burson’s photos of chimaeras.
