When you open your Facebook, when you look through your images, make a search on google, a transaction on your bank account… An algorithm is there, to bring it for you, attempting to find fraudulent and to optimize your preferences.
In the past, human build algorithms bots by giving them a particular set of instructions that human could explain. ‘If that, then do this’. But many of those problems became too complicated and abstract in order for humans to write simple instructions for a computer to follow.
It happened to be, that computer vision, aka, automated image analysis, is one of those problems.
The inspection industry is in a race to carry out atomization process that is accurate enough, to go through images and quickly classify the images into their correct classification.
Few months into 2018, and we can take an educated guess, despite lots of promises, there is no commercially ready algorithm that can do just that, yet (and we would love to know if we are wrong).
Let us take a dive into what is the current situation on the market, how far are we from getting there, and why, ‘the f***’, is it a game changer?
Read the full story here >>