My phone can recognise faces, what do I look like to my phone? This prompted an investigation of Computer vision algorithms and face recognition. I created an android app which captures or loads an image, processes it and publishes it on twitter: a selfie, a self-portrait, an auto-portrait, an autoselfie. The android app becomes the interaction mechanism. The twitter hashtag #autoselfie becomes the exhibition space.
Can people recognise us from our selfies on twitter? Can software see the faces too? Do we want to be recognisable by humans and/or computers? There are forms of identity obscuring which covers everything but the eyes. There are other forms which only cover the eyes.
The selfie you post to a social network site traverses layers of software. Do you know what the software does? Can you see what the software does? This app takes an images suggested by the user/viewer and modifies it before it posts it. The user has less control. The resulting image can no longer be recognised as a face by algorithms reading the feeds. Can the human viewers of the feed still recognise you?
The code for this app is visible but possibly not accessible. This app gives me a platform to play with images, to let other people try my code and to make it visible to everyone on twitter.
Accompanying this project there is record in a blog called #autoselfie. At the moment it is documenting a series of remix images created with #autoselfie app. The images are taken from the 2013 SPI national portrait awards held in South Africa. The SPI portrait awards excluded lense based works and all the works show the identity of the sitter as interpreted by the artist. In many cases #autoselfie app recognises the faces.