Computer Vision News - April 2016

Guest Paolo Raschiatore - Vision Device Many professional have dedicated the work of a lifetime to the field of computer vision, image processing and machine vision. Some of them did it in the academy, others in the business arena. Their vision and thoughts are often original and Paolo Raschiatore, CEO of Vision Device , has kindly accepted to be our guest in this first issue of Computer Vision News. Paolo Raschiatore has entered the field of machine vision back in 1982! As you will see, he does not only have a lot to say about the past; he has also plenty of advice for the future… COMPUTER VISION NEWS Computer Vision News: W hat do you think has changed in our field since 1982? Raschiatore: Many things have changed since. At the time, Personal Computers and digital cameras did not exist; huge hardware manufacturers which we know today were not born yet. All software and hardware had to be done in-house. We started from electronic boards designed to capture images, since existing computers were not powerful enough to process images. CVN: Were you aware of being pioneers in a new field? Raschiatore: Probably yes: a software developer’s initial tasks was usually to program electronic circuits; we even had to build the assembler. It was certainly a pioneering experience. CVN: Which technical innovation surprised you the most in the last 30 years? Raschiatore: That may sound like a paradox, but methodologies remained pretty much the same: though they gained in sophistication, progress was always achieved in moderate, progressive steps. What dramatically changed is the experience we have acquired on the way, generating applications which today fascinate me: on one side 3D solutions , which recognize objects in all their dimensions and forms, even when the image shows them mixed with other items; on the other hand, Artificial Intelligence used in the vision field, with neural networks and the like. Algorithms learning to sort out defective parts in production lines are a very powerful solution in the industrial domain. I am also impressed by automatic interpretation of images, like when software is able to automatically understand and interpret the scene it observes. All over the years, we have often dealt with the difficulty of teaching a program to extract the main features of a scene the way our brain does. But during the last year or so, it has become frequent to read about models which seem to offer a working solution to this challenge. 6 “ Be open-minded, observe what is done in all fields ”