What is
computer vision
and how does
it work?

Computer vision is a concept that has been around for several decades and, as technology in general has advanced, so too has computer vision. But what is computer vision and how does it work? 

Computer vision is a field of artificial intelligence (AI). It enables computer systems to capture digital images, video and other visual input and then derive meaning from that data, so it can make recommendations based on the captured information.

Computer vision is not just about attaching cameras to computers to allow them to "see" things—it also enables them to actively observe and understand what they are seeing.

A key facet of computer vision is that it generates a great deal of data that must be sent between the camera and the compute engine. Given the large volume of data that is required to be transferred from the cameras to the compute element and given the fact that this data transfer needs to happen in as near as real time as possible, 5G networking at the edge would be a great candidate technology to support this type of application.

Here's a list of the core elements needed for computer vision:

• a device (such as a camera or other type of sensor) that acquires input
• a power source
• a processor
• a minimum amount of onboard data storage
• the networking capabilities to transmit that data

Pattern recognition algorithms allow a computer vision system to mimic the human brain and understand what it sees. With the help of machine learning (ML) and deep learning methods, it can recognize a variety of visual information.

Having answered the question "what is computer vision and how does it work,” let's look at its applications. Computer vision technology must be accurate; in some applications, it can mean the difference between life or death.

Here are a few industry-specific examples and the requisite capabilities for computer vision success:

• Quality control: Computer vision has the potential to reduce product defects because it's continuous and highly accurate. It can also contribute real-time information and insights from the factory floor—a hallmark of Industry 4.0—as well as improve safety.
• Medical imaging and health: Here, it could potentially assess X-rays and scans more accurately and faster than a medical specialist, while AI can potentially identify rare diseases not every human doctor is familiar with.
• Agriculture: Automation and detection capabilities can help to make yield analysis more accurate by collecting data from drones and satellite images and combining it with environmental data, such as soil factors, weather and moisture conditions. It can also monitor livestock.
• Shopping experiences: Retailers can help improve shopping experiences using virtual mirrors and recommendation engines so shoppers can “try on” clothes and accessories using mirrors equipped with a special display behind the glass.
• Automotive: Computer vision collects data from vehicles and infrastructure to model driving behavior, improving autonomous driving performance and vehicular safety through such functions as lane finding, for example. Computer vision can be used to identify pedestrians and vehicles and inform both human drivers and autonomous driving systems in real time, when using reliable and secure 5G networking at the edge.