Last month, I had the pleasure of attending this year's Vision Show that was held in Boston, Massachusetts, USA from May 3-5. There were many interesting products, events, technologies and presentations at the show. Notably, Steve Varga of The Procter & Gamble Company gave a brilliant keynote speech that highlighted the perils of being a systems integrator. What I will especially remember is the joke he first told at the beginning of his presentation. Unfortunately, my publisher, Alan Bergstein did not want me to relate this story to you. If you were there, you would know why.
Another interesting event at the show was the "Shark Tank"-style Startup competition where companies from all over the world presented their ideas. Eyes of Things, an EC initiative, were there to describe how they were going to build the world's smallest vision system. A company called Skye Automation showed a great video on how to sort categories of lumber, SUALAB showed how they were using deep learning to detect defects on leather car seats and Suspect Labs gave a presentation on face verification.
But the company that won the $10,000 prize was a company called Brohan that had used image processing algorithms to analyze videos and automatically drop ads into flat surfaces on the frames. I was rather surprised at the judges' decision until I realized that the criterion to win was whether by investing in such a company, you would make a lot of money.
Trends on the show floor could be summarized as those dealing with (a) interface standards (b) deep learning and (c) miniaturization. Now, systems integrators can choose standards such as USB, Ethernet, GigE, GigE Vision, 10Gig, Camera Link, CoaXPress, Camera Link HS and upcoming interfaces such as Apple's Thunderbolt and NBASE-T, that have already or will implemented by companies such as Ximea and Pleora. What was also interesting to see was the number of camera and frame grabber companies that have adopted fiber interfaces for some of these standards.
A number of companies also showed systems that combine traditional image processing algorithms such as edge detection, color and blob analysis with classifiers such as SVMs, k-Nearest neighbor and neural networks to perform image classification. Alternatively, some systems were shown where images were presented to such classifiers without the need to perform pre-processing.
While the Eyes of Things EC project aims to develop the world's smallest vision system, commercial vendors may beat them to it. Microscan - one of our 2016 Innovators Awards honorees-showed me its latest MicroHAWK camera that cost about $500 (without software) that was about the size of half a regular cigarette packet.
Although it is impossible to summarize all of the products, events and technology at The Vision Show in a single editorial, atwww.vision-systems.com, you can see a video review of the show.
|Andy Wilson, Editor in Chief|