With the advent of social media and insta-messaging apps like Facebook, Whatsapp, Instagram and Snapchat there are as many as 22,000 pictures, as per Photoworld, being uploaded each second.
The first camera phone was introduced back in 1999, and now, mobile phones sport dual cameras. With the advent of social media and insta-messaging apps like Facebook, Whatsapp, Instagram and Snapchat there are as many as 22,000 pictures, as per Photoworld, being uploaded each second. Thus far, the functionality of the camera has been restricted to just clicking images or recording videos. But a new generation of artificial intelligence software and algorithms are giving smartphones the ability to not only take pictures but understand and differentiate between objects in pictures, much like humans do. Recently, Google acquired a French image recognition start-up called Moodstocks just weeks after Twitter acquired Magic Pony to incorporate machine learning into its processes. The acquisition comes just a couple of months after Google had announced messaging apps that can recognise and catalogue images.
While image recognition can certainly extend the world of photo-searches and is a great tool for user engagement given that there is a large database to learn from, its real application lies in the bio-medical field. It is being used by companies like IBM to assist doctors by processing medical images to help diagnose diseases in a faster manner, while there are also experiments going on to develop devices for visually impaired which can provide voice navigation based on image recognition. The technology is also expected to benefit augmented reality devices, by providing instant access to information around us and also making for smarter robots and self-driving vehicles. As for privacy concerns, especially regarding image recognition, we should not fret too much as we have already shared our lives on social media—all that image recognition is doing is to learn from it and build on it.