In the future, the vast majority of photos and videos recorded won’t be seen and used by humans – they’ll be seen and used by machines. This week we interview Allan Benchetrit, CEO at Algolux – a Montreal-based AI company focusing on computational imaging.

If you take an image for a human being in a consumer application (maybe an iPhone app or a recreational DSLR camera), you probably want it to be visually appealing and clear to the human eye.

As it turns out, machines don’t need pretty images, they need to do their jobs. If a computer vision system needs to detect road signs, or suspicious people in an airport, or the presence of weeds in a cornfield – it may create images that are ugly to the human eye, but perfectly calibrated for being interpreted by machines for their jobs. As it turns out, this is a complicated AI-related problem itself, and Allan walks us through it.

If your business uses cameras heavily – or may do so in the future – this interview will provide an around-the-corner look at what it takes to create effective computer vision applications.

Listen to the full interview below on Soundcloud:

itunes-podcast
itunes-podcast
stitcher-podcast
google-podcast

Guest: Allan Benchetrit, CEO at Algolux

Expertise: Computer vision, computational imaging

Brief recognition: Allan holds and MBA from the John Molson School of Business at Concordia University. Before Algolux, he served as CEO at mobile video company Vantrix. Before Vantrix, Allan was VP of Sales at Wysdom, a software company based in Montreal.

Big Idea

Soon, most video recording will exist for machines to make sense of, not humans.

Interview Highlights with Allan Benchetrit of Algolux

Listed below are some of the main questions that were posed to Allan throughout the interview. Listeners can use the embedded podcast player (at the top of this post) to jump ahead to sections that they might be interested in:

  • (3:15) How does “sensor tuning” within a camera actually work?
  • (7:00) What actual elements are being adjusted when sensors are being tuned for a new camera application?
  • (14:40) How might vision systems need to be tuned to fit the specific use-cases just for humans, or just for machines?
  • (21:00) For autonomous vehicles, what are the most important camera capabilities for computer vision?
  • (19:20) What the the problems with cameras today – what paradigm shift is necessary to get to more dependable computer vision?

Other Computer Vision Interviews at TechEmergence

Our AI in Industry podcast (available on iTunes and Soundcloud) is one of our main content efforts here at TechEmergence. Each week we interview AI researchers and executives to learn about the applications and implications of AI in organizations today. If you enjoyed our interview with Allan this week, you may enjoy some of these other interviews focused on computer vision:

MARKET RESEARCH x INDUSTRY TRENDS

TechEmergence conducts direct interviews and consensus analysis with leading experts in machine learning and artificial intelligence. Stay ahead with of the industry with charts, figures, and insights from our unparalleled network, including executives from Facebook, Google, Baidu, Yahoo!, MIT, Stanford and beyond: