Graphics in reverse

I’m just wrapping up a fun and challenging consulting project in image processing and computer vision. A big part of the fun was the team: I enjoyed working with them, and teaching the methods underneath the work as we went.

I came to computer vision by way of computer graphics, and the two are the same problem run in opposite directions: graphics turns a description of a world into pixels, and vision turns pixels back into a description. A lot of the best ideas cross over. Graphics engines learned long ago not to spend effort evenly, drawing what is near in fine detail and what is far coarsely, and the same instinct pays off in vision: put the computation where the information is.

The client’s Chief Technology Officer was kind enough to write this:

Working with David at Lixel LLC has been a highly professional experience. He brings deep industry experience and a rare ability to turn complex analytical and statistical processes into efficient, scalable, and elegant solutions.

David is building an imagery-analysis platform for our team while also helping us understand the underlying fundamentals and methodology. His approach makes sophisticated work feel clear and approachable, an indicator of real technical mastery. He has been an excellent partner, a generous teacher, and someone who has inspired confidence and enthusiasm across the team.