“I dreamed about a human being” is part of a project exploring the use of artificial intelligence as applied to photography by using online open source code and data. The project already has a database of 56 million images. We have freely accessible amazing tools and databases of gigantic images, but have not yet fully understood what we can do with them or what it means that they are there.

This series of images is the result of statistical calculations on 257 faces detected by an algorithm that has been taught to recognise portraits of a specific aesthetic quality. The search was conducted over 5 million images with Creative Commons licenses posted on Flickr. “Mean”, the displayed image bellow, is the average of these 257 photographs, comprising 17 babies, 106 men, 79 women, 18 girls, 23 boys and 14 errors.

Another way of viewing these accumulated images is by seeing them pass by at full speed and blurring your vision a little. You can see the same “average effect” by watching the following video, which has 2,582 images moving at a rate of 25 images per second http://vimeo.com/49552899

Photography and artificial intelligence

“I dreamed about a human being” is part of a larger project exploring the use of artificial intelligence as applied to photography by using online open source code and data.

The project began in 2008 and it has a database of 56 million images with Creative Commons licenses. In 2011 the image search ended in order to begin processing the images. In later versions the system was able to add 200,000 photos per day to the database.

Ten years ago having access to these amounts of information and being able to process them on a home computer with artificial intelligence algorithms might have been considered science fiction.

The phenomenon that has led us here today involves three dimensions: digital cameras and social networks, the philosophy of open source sharing and the increased capacity of computers and networks that can be summarised in the popularisation of digital technology.

Digital cameras and social networks encourage the existence of a large number of captured and published images. These images can be accessed through public programming interfaces, which allows you to programme computers to access these images almost limitlessly.

The concepts of open source and its ideology currently no longer belong to the world of software and have expanded into usage licenses for almost any type of content. Consequently, users allow access and explicitly grant permission for reuse by publishing these images with Creative Commons licenses.

At the same time, many companies and universities have realised that releasing the code for certain parts of their research can help them sell products or develop projects with input from the open source community. As a result, quality open source exists today specialising in computer vision and artificial intelligence. The technology used in this project was developed and released by Intel, Compaq and Mitsubishi.

All this can be added to increased computing capacity and the transfer speed on the Internet.

We are all getting into digital technology as we all got into cars, although it took 100 years to see the impact on the environment. But the impact of this technology does not end up in the atmosphere; it is entering our brains and even altering its structure. It is essential that we at least try to understand the capabilities of the technologies that we are using.