We live in a sea of faces. It’s not just that the planet currently houses the highest number of faces in human history (7.3 billion and counting). Social media has effectively transformed our faces into our digital calling cards, representing us in the increasingly hazy space between our real life and Internet existence.
So what would happen if we morphed 1,000 of those unique faces into one single image: The Face of Toronto? On October 3, we’ll answer that question.
We’re thrilled to introduce Architech Labs to the public for the very first time at this year’s Scotiabank Nuit Blanche. At the annual all-night celebration of art and culture, we’ll showcase how some of the cutting-edge facial analytics technologies we’ve concocted in our Labs can capture, process, and combine countless faces in mere seconds.
#TheFaceofTO installation began as a simple idea by Labs junior developer, Jin Sung Kang. He wondered what the collective face of one of the world’s most diverse cities would look like if a large sampling of those faces were morphed together at once, in real-time.
While this sounds like a simple enough task for today’s computers to perform, we knew we’d only have a few seconds to process each face for the exhibit to work. We took this problem to the rest of our lab geniuses, who’ve spent many months building out the incredibly advanced deep learning models that can now recognize individual keypoints with great speed and robustness in uncontrolled conditions, similar to the human brain.
That means people of all genders, ages, and backgrounds who arrive at our Nuit Blanche exhibit can have their photo taken and watch as their face is immediately projected onto a gigantic screen. From there, our deep learning computers automatically detect the keypoints of each face and morph them into the hundreds of other faces that have already passed through our exhibit that night.
Every face will change the composite photo as the evening progresses. At the end of the night, we’ll discover what Architech’s Face of Toronto will look like. How will our deep learning computers perceive, interpret and visualize what it sees? Your guess is as good as ours. Come downtown and find us Saturday night at the corner of Adelaide and York and make your contribution to #TheFaceofTO!