Herpes a la Face Tracking

For my final in Nature of Code and Computational Cameras, I worked on creating an app that would allow users to take pictures of themselves or their friends. The image is scanned for a face. If a face is detected, simple rules of proportion are used to determine the basic location of the mouth. After the mouth is found, a script that mimics the look and feel of oral herpes is applied to that area of the picture. The resultant image is gross and funny. I also created another version that turns a portrait picture into a mugshot, includes criminal information and a lone teardrop tattoo. Recently I have been doing a lot of image augmentation. I think that it could prove to be very popular if I could only get it out to the masses. Further ideas include, fake family vacation pictures and fake historical pictures. It seems that everyone loves neat pictures of themselves, and frankly, they don't even have to be real.

Camera Futura

I was asked to think about the future of cameras and how they will impact or lives. I came up with a few ideas, some more distant then others.

Short Term
I think image stabilization could easily be improved by allowing the user to hold the shutter button down, but only firing the shutter when an accelerometer in the camera says that the camera is appropriately stable.

Medium Term
Through advances in image pattern recognition and optimization, new picture taking modes could exist that search for classical balance and proportion. Only when a composition presents itself that includes 20 golden rectangles can an image be captured. This would help amateur photographers better understand the principles of composition.

Long Term
The ability to compare dimensions, specifically in relation to your body, seems like a useful piece of information to be included in the photo. Which makes me wonder of the future of the photograph itself. In the future cameras will most assuredly capture information in a three dimensional format. Probably through the use of stereoscopic vision like our own eyes. This would weave the extra information into the display rather than simply including it in the Metadata.

Color Tracking Fun! part 3

Back at it with color tracking and the gang. This time are sights are set on driving the streets of NYC in a virtual car. We had several immediate problems that made it impossible to work outside. It began raining and we were using an auto gain webcam which was making our lovely neon pink squares appear white.

We grabbed a better camera and headed to upper floors of Tisch to shoot in a larger location. The code was riddled with problems and made it nearly impossible to get satisfactory shots. Nevertheless, Winslow had an idea brewing and was able to stitch together a story of a class field trip.


Wicked Awesome!!! from Winslow Porter on Vimeo.

Elffervescence made a debut in the space scene, this time wearing a black background.

Angela's awesome documentation again.

Color Tracking Fun! part 2

The following week we all met again and decided to have a dance party. We invited some of our friends and danced in front of a short throw projector projecting party videos from amongst the ISH. Music by Mylo, "I'm back".


MASKPARTY featuring Coconut Horse, ODB, Lemonhead, and Mario from Winslow Porter on Vimeo.

This time the code was changed to calculate the angle between the two green squares and display the images of various heads onto our bodies. We created a make shift head gear device out of black foam and rubber bands, to which the green squares were attached.

Another inside look by Angela Chen.

Color Tracking Fun!

After working with blobs and the OpenCV library to create "Granpa Loves Valentine's Day", Winslow Porter and I combined forces with Meredith Hasson and Angela Chen to tackle color tracking.

Using Dan O’Sullivan's color tracking code as a starting point, we began to develop a simple and sometimes reliable way of determining the location of two green squares from the pixel data of a webcam. From the location of each square we can then determine the distance between the squares and even the angle of a line drawn between the two squares.


Gone Fishing Debugging from Matt Richard on Vimeo.

But what to do with such a thing? After a quick brainstorm, the act of fish-story telling, specifically the phrase, "a [fish] this big!" would be captured in augmented video and a simple story would be told.


Catch of the Day from Winslow Porter on Vimeo.

The code can be found here along with an awesome inside look from Angela Chen.

Granpa looks for OpenCV in the kitchen

Winslow Porter and I played with the new OpenCV library for Processing. Thankfully Granpa was around in the kitchen for us to Capture.


Grandpa Loves Valentines Day from Winslow Porter on Vimeo.

We also found footage from 1969 of a van making an illegal U-turn in San Francisco.


Rainbow Image Detection from Winslow Porter on Vimeo.
This effect was created by Andy Best in a tutorial he wrote for processing's OpenCV library. You can check out the tutorial here.

Then I stayed up playing with the sketch until I went to class. I did capture an amazing smile though.


All-nighter smile from Matt Richard on Vimeo.

Pictures from early on in the evening.

Hue is New York?

I am interested in using cameras to gather color data about New Yorkers' wardrobes. The information could be used to create the ideal outfit for a specific part of town or to create a wonderful portrait of the city.