Back to articles

Top 5 Disruptive Computer Vision Applications

25th February 2019
By Coline Chauffard
Community Executive
Machine Learning & Deep Learning

Computer vision is the technology that systems use to understand the images they are capturing through image sensors. With the increased use of this technology due to the rapidly decreasing costs in processing speeds required for this technology to flourish. Computer vision is set to fundamentally change the ways we use and interact with our environments. In this article, I outline what computer vision is, why businesses are using it and what the current top 5 ways computer vision is disrupting the industries it is applied to.

Disruptive Applications of Computer Vision

What is computer vision?

Computer vision is a field of artificial intelligence that teaches computers how to understand images.
Our brains are very complex when it comes to deriving meaning and motion using our eyes, and teaching a computer how to do the same is actually a complex problem. However, in recent years computer scientists have begun to crack some of the challenges of computer vision, and the implications for everyday life are astounding.

Computer vision involves three basic challenges:

  1. See – Seeing an object is a nontrivial challenge. We have to teach computers how to see distinct objects in a photo or video and separate those objects from the environment around them.
  2.  
  3. Describe – This involves categorising, labelling, or comparing the objects in a photo with other objects the computer has seen in the past.
  4.  
  5. Understand – Finally, we need to teach the computer what to do with the information it gains from an image. For example, in the case of self-driving cars, the computer vision needs to disregard advertisements, but it should obey street signs. Teaching the computer the difference between the two is vital.

Advances in machine learning allow researchers to feed databases of images into learning algorithms. The result is we no longer need to describe to computers what they see. Instead, they can learn to interpret images over the course of many trials.

Computer vision can assist humans to do jobs better, whether that’s surgery or security. It can also process thousands of images at once, providing rich metadata and insights about collections of images as a whole. Finally, it never gets tired and can be fairly unintrusive, opening up a host of applications for observing, securing, and gathering real-time data about a space.

How and Why Businesses Are Using Computer Vision

Companies are quickly recognising the implications of the disruptive applications of computer vision, and many top companies have invested in computer vision. In fact, 2016 saw capital funding for these types of artificial intelligence projects rise to £398 million ($522 million), a new high. Funding for such projects doesn’t seem to be slowing down.

In recent years industries have discovered ways to leverage this technology for business cases. The potential applications could change everything from our doctor’s visits to the way we shop. Companies are using computer vision to personalise their marketing, drive conversions, increase customer satisfaction, and safeguard sensitive information.

The top industries primed for disruption:

1. Computer Vision Applications in Advertising/Marketing

Computer vision has the potential to revolutionise advertising by personalising messages based on their consumer behaviour. In the world of offline advertising, for instance, companies are already experimenting with tracking facial features and eye movement of passers-by. This gives advertisers data about how their messages are performing with audiences, including how long people look at the advertisement and how they react to the content. Computer vision could make personalising online video content easier as well by adding metadata to images so that advertisers can serve many varieties of the same ad based on the viewer.

A rather obtrusive example of computer vision being used in advertising is highlighted in the 2002 movie Minority Report.

Beyond video content, companies are looking for ways to incorporate advertising into VR, and computer vision is a part of detecting and filling spaces where ads could be placed in VR and augmented reality. New 3D-capable computer vision projects might make it possible to create 3D and VR-ready content directly from a phone or camera, leading to an explosion in immersive storytelling and marketing.

2. Computer Vision applications in Retail

In the face of rising pressure from e-commerce, brick-and-mortar retail has stepped up the hunt for new technology to enhance the in-person shopping experience. One application involves using computer vision to analyse the baskets of in-store shoppers. Based on the contents, retailers can recommend other products the shopper might like or send an employee to assist customers who seem to be having trouble.

Of course, online retailers are also looking for ways to leverage computer vision. eBay is looking into technology that allows users to take a picture of an item they like and eBay will serve them similar items for sale. Pinterest has taken the same concept a step farther with Pinterest Lens. The software allows a user to take a photo, find similar products, and the algorithm learns the user’s tastes over time.

At the intersection of brick-and-mortar and e-commerce is Amazon Go, the shop with no tills. Customers can take items from the shelf, and a combination of computer vision and other sensors will detect what item they’ve taken and automatically add it to their online baskets. When the customer is finished shopping, they simply walk out of the store, no need to queue.

Further reading – Welcome to the store of the future

3. Computer Vision Applications in Healthcare

Computer vision is currently being deployed against microscopy, X-ray, angiography, ultrasound, and tomography images to reduce white noise and help doctors pinpoint potential problems. Similar algorithms are in use in dermatology, where computer vision is helping identify potential skin disease or skin cancer.

Another healthcare application of computer vision is a surgery guidance tool currently in development that would help surgeons make decisions during laparoscopic surgeries.

Computer vision will also be used in medical research. Using anonymized scans and images from past patients, researchers can use computer vision to discover patterns in disease progression.

Further reading – 4 Examples of Computer Vision and NLP in Healthcare

4. Computer Vision Applications in Transport

Self-driving cars are among the hottest developing technologies in the world and for good reason. Computer vision can greatly reduce vehicle crashes, injuries, and fatalities. Most new cars currently implement computer vision for lane keeping and obstacle detection. In the future, computer vision will play an essential role in fully autonomous driving.

Currently, computer vision is not 100% reliable for fully autonomous driving. However, most experts agree that the technology is only a few years away from being better at driving than humans are.

5. Application of Computer Vision in Biometrics

Biometrics is the process of using biological markers to confirm identity. Perhaps the most common example of a biometric device is the fingerprint scanner in your smartphone. It grants access to your information based on your biology.

Computer vision is used to develop touchless forms of biometrics. For example, Facebook’s facial recognition software that asks if you want to tag a specific friend in a picture is an application of using computer vision to determine identity. However, sometimes Facebook’s computer vision makes a mistake and asks you to tag the wrong friend. When it comes to security and access control, there’s no margin for such mistakes.

The new iPhone X, some laptops, and other new phones can use facial recognition to grant access, and several banks, like HSBC and Schwab, are pioneering voice access for account information over the phone. As these technologies continue to develop, expect biometrics to play a growing role in everyday identity verification.

The latest research in the field of computer vision in biometrics involves a concept of “determined liveness.” In the future, your face may serve as your password, automatically granting access without needing to type. But in order to prevent an attacker from simply holding up a picture of your face, determined liveness involves taking several photos in rapid succession to verify you along with other means of biometric verification.

Computer Vision’s Disruptive Impact

These are just a few of the industries that computer vision will disrupt. The technology also has far-reaching implications for hospitality, education, agriculture, manufacturing, and many other industries. Computer vision technology is still very young. As funding for computer vision projects increases, we’ll see new breakthroughs in computers’ ability to interpret visual information, leading to an explosion in the way we work, travel, and live.

Keep reading:

Computer Vision in Retail

4 Examples of Computer Vision and NLP in Healthcare

8 key ethical dilemmas we face in artificial intelligence

The ways AI is transforming healthcare

Is your company currently looking to implement computer vision? Are you working in this field? I would love to speak to you about the projects you are working on. Please get in touch with me at [email protected]

Follow me on twitter: @SteveAtLogikk

Share Article

Get our latest articles and insight straight to your inbox

Hiring data professionals?

We engage exceptional humans for companies looking to unlock the potential of their data

Upload your CV
One of our consultants will contact you
to find out more about you