Mapping startup hires Apple computer vision expert

“Mapillary, a Malmö, Sweden-based map technology startup, has hired a computer vision expert away from Apple Inc, the company said Wednesday,” Stephen Nellis reports for Reuters. “Till Quack, who worked at Apple as an engineering manager overseeing augmented reality and self-driving efforts, will become Mapillary’s vice president of product, the company told Reuters. ”

“Mapillary was founded after CEO Jan Erik Solem himself left Apple in 2013 after selling his facial recognition startup, Polar Rose, to the iPhone maker in 2010,” Nellis reports. “Mapillary aims to solve one of the most expensive problems in mapping: keeping maps up to date with so-called ‘street level data’ about street signs, addresses and other information that can be observed from the road.”

“The way big companies such as Alphabet Inc’s Google and Apple Inc solve the problem is by sending out fleets of vehicles outfitted with cameras and other sensors to gather images of the roads and keep maps updated,” Nellis reports. “Mapillary’s approach is to take software that can ingest pictures from nearly any kind of camera, such as a smart phone or a GoPro action camera, and use so-called ‘computer vision’ technology to stitch that together into a three-dimensional map. The company provides that capability to mapping firms, such as HERE Technologies, the mapping firm owned by a consortium of German automakers.”

Read more in the full article here.

MacDailyNews Take: A bit of free pub for a mapping startup. Hopefully, something useful comes out of it for all of us.


  1. I’m sure a specialized mapping vehicle would be much more efficient than some random photo captures. I’d think it would be best to keep all the images at the same resolution. Both Google and Bing have been creating and updating their maps with specialized mapping vehicles for years. It’s unlikely they’ll be dumping their fleet of mapping vehicles any time soon.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.