Learning To See – Making deep neural network predictions on live camera input

Author Filip VisnjicCategory Members openFrameworksPosted on: People

Latest in the series of experiments and explorations into neural networks by Memo Akten is a pre-trained deep neural network able to make predictions on live camera input – trying to make sense of what it sees, in context of what it’s seen before.

Learning To See is an ongoing series of works that use state-of-the-art Machine Learning algorithms as a means of reflecting on ourselves and how we make sense of the world. The picture we see in our conscious minds is not a direct representation of the outside world, or of what our senses deliver, but is of a simulated world, reconstructed based on our expectations and prior beliefs. Artificial neural networks loosely inspired by our own visual cortex look through surveillance cameras and try to make sense of what they are seeing. Of course they can see only what they already know. Just like us.

Sorry, this is Members Only content. Please Log-in.

Join us today by becoming a Member.

  • Archive: More than 3,500 project profiles, scores of essays, interviews and reviews.
  • Publish: Post your projects, events, announcements.
  • No Ads: No advertisements, miners, banners.
  • Education: Tutorials (beginners and advanced) with code examples, downloads.
  • Jobs Archive: Find employers who have recruited here in the past (over 800 jobs).
  • Discounts: Special offers and giveaways (events, books and media).

Become a Member

/++

Trace Ribbon – An emergence by movement

Created by Shohei Fujimoto, Trace Ribbon is a device that automatically and continuously records and plays back movement. From reading the movements of the user via Leap Motion, it mirrors an organism that does not actually exist while simultaneously gaining an understanding of the rules of the movements that are physically taking place.

Bright Matter: Afterglow – Joanie Lemercier on light, craft, and the sublime

From November to January 2014, Muriel Guépin Gallery in NY was the home to ‘Bright Matter’, an exhibition of enigmatic works by five international artists widely recognized for their spatial-aesthetic research, creative engineering, and stunning perceptual hacks. We check in with the event’s instigator, curator, and participating artist Joanie Lemercier for a report, the back story, and a 2015 teaser.

Z1FFER – A True Random Number Generator for Arduino (and the post-Snowden era)

Low cost and open source, Rob Seward’s Z1FFER is Hardware Random Number Generator that harnesses thermal noise to provide a high quality random bitstream for research and experimentation. Available at the CAN shop now!

Kinedioscope: Muir Woods – Advancing photogrammetry for depth and masking

Created by London based convivial studio, Kinedioscope is a technique used to create animated depth effects on static photographs. The process is comprised of reverse-engineering the technology of photogrammetry in order to perfectly align the photography with the perspective of the 3D model and create depth and masking effects.

/+

/++

Laser Projection Mapping on Soap Bubbles by Memo Akten

What happens when you mix soap bubbles and lasers? Memo Akten, 1/3 of Marshmallow Laser Feast, just got hold of Etherdream DAC laser and has been “messing” with it using openFrameworks.