Foliage waving in front of the camera, sunglasses, goggles or even a full-face helmet cause no problems for the intelligent face-recognition and tracking abilities of the Canon EOS R5's autofocus. Taken on a Canon EOS R5 with a Canon RF 28-70mm F2L USM lens at 30mm, 1/3200 sec, f/2.2 and ISO100. © Félicia Sisco
How does Canon's intelligent autofocus work, which Canon cameras use deep learning AF, and how can this latest Canon autofocus technology help you to capture better shots and improve your hit rate?
Originally developed for the Canon EOS-1D X Mark III, the intelligent autofocus system with deep learning algorithms is also present on the Canon EOS R5 and Canon EOS R6. Wedding, fashion and portrait photographer and Canon Ambassador Félicia Sisco has already found it invaluable.
"It was crazy, with wedding guests jumping up and down all over the place," Félicia says of a recent assignment. Such a scenario might sound like an autofocus nightmare, but Félicia's Canon EOS R5 never missed a beat. In fact, she has been impressed with the EOS R5's AF since her first experience with it. "The first time I used the EOS R5 was for a very difficult fashion shoot with lots of movement," she says. "The autofocus was incredible. Now I'd feel lost without it." Félicia has also used the Canon EOS R6, which utilises the same deep-learning artificial intelligence autofocus system.
Canon Europe Professional Imaging Product Specialist Mike Burnhill explains that this latest iteration of the EOS iTR AFX system uses 'deep learned' artificial intelligence. The system is based on an algorithm that teaches itself by scanning millions of images. The system essentially learns how to recognise the heads of people, even if they're skiers wearing goggles, racing drivers wearing helmets, or gymnasts upside down or even facing away from the camera. Deep learning is a huge leap forward from the intelligent autofocus systems of previous Canon cameras.