Tech Trends in Practice

TREND 12Computer Vision and Facial Recognition

The One-Sentence Definition

Computer vision, also referred to as machine vision, is where machines (including computers, software, and algorithms) can “see” and interpret the world around them – with facial recognition (which uses computer vision to identify people) being a prime example.

What Is Computer Vision and Facial Recognition?

Early experiments in computer vision began as far back as the 1950s, and the technology was already being used commercially by the 1970s to interpret typed and handwritten text.¹ So, if it’s not a new technology, why highlight it as a key trend today? To answer that question, we first need to get a quick (and non-technical) explanation of how computer vision works.

As a form of artificial intelligence (AI, see Trend 1), computer vision is essentially all about processing, analyzing, and making sense of data – it’s just that the data being analyzed is visual rather than, say, textual or numerical. For the most part, this means the data being analyzed is in the form of photos or videos, but it could also include data from thermal and infrared cameras and other visual sources.

Analyzing visual data with any real accuracy relies on deep learning and neural networks (Trend 1) – in other words, using pattern recognition to distinguish what’s in an image, after learning from a specific data set of other relevant images. For example, in 2012, Google was using a neural network to identify cat videos on ...