Have you ever felt overwhelmed by the appeal and beauty of an artwork, but disappointed by the scarcity of information available on the label next to it? Have you ever thought of your smartphone as a way out?
The applications developed by Project ARM rely on the functions of Image Recognition (IR) to provide you with a unique experience. Thanks to such technology, these applications are able to recognize an artwork and give you access to a database of related multimedia content. How that? Just by pointing your smartphone toward the artwork in question, be it a painting or a sculpture.
But what is Image Recognition and what do we mean by saying “recognize”?
What is Image Recognition?
The way our brain works makes the action of recognizing objects quite straightforward. We experience no difficulty in our daily life when it comes, for instance, to identifying a painting or a sculpture and distinguishing one from the other, thus labelling them as belonging to the “painting” or “sculpture” category. Actually, attempts at replicating the natural system that allows the recognition of an image (and its subsequent classification) by means of IT equipment have shown, instead, how infinitely complex this system is. However, over the last years, a branch of IT known as computer vision has made huge progress in this respect.
Nowadays, thanks to the implementation of algorithms and methodologies allowing the recognition of shapes, colors and moving objects, a wide spectrum of tools has been introduced that makes it possible to identify and classify the objects and figures that populate our world.
A timeline of Image Recognition
Having become more familiar with the concept of Image Recognition, let us get into the detail of how this specific technology emerged and evolved. Here below, we drew up a brief timeline of the main stages it went through:
2001 : Paul Viola and Michael Jones invented a simultaneous face detection algorithm allowing for human figures to be identified through their facial traits.
2005 : Navneet Dalal and Bill Triggs published Histograms of Oriented Gradients (HOG), theorizing a feature detector for the recognition of pedestrians in security system circuits.
2012: Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton hit it big with a new object recognition algorithm ensuring an 85% level of accuracy.
2015: the Convolutional Neural Network (CNN) developed IR tools whose level of accuracy in facial recognition exceeded 95%.
Today: Google, Amazon and even some car manufacturers are channeling their efforts as well as their R&D investments into the development of new technologies that integrate Image Recognition.
Amazon, for instance, recently launched Amazon Rekognition, a new product based on Image Recognition functions and able to scan photos, guess at emotions through face recognition and classify objects or animals.
Welcome to the future
As you might have guessed, within few years or maybe months, Image Recognition (IR) will turn into a useful ally in our daily life. Are you ready to make the most of it?
Let us slip in some advice: start training right away, with the image recognition applications by Project ARM! Scan and recognize your favorite artworks and set out on a journey into the secrets of art.