Computer Vision: Keeping an Eye on You

Sam Brown our consultant managing the role
Author: Sam Brown
Posting date: 5/9/2019 3:38 PM
5
Animal Farm, Fahrenheit 451, and 1984 all wove tales of Big Brother watching. Remember when only Superman had X-ray vision and leapt tall buildings in a single bound? Well, as big tech grapples with the cause and effect of its place when it comes to facial recognition, security, and object detection, there’s a new superhero in town. Computer Vision. 

The Eyes Which Sweep Our World


While we still can’t be everywhere or see everything at once, we have eyes everywhere. Despite security cameras, motion detectors, satellites, computers, and smartphones, there are still dangers we may miss.

So, by making our machines and computers visually-enabled using Artificial Intelligence, we significantly ramp up what we can see and how fast we can process it. Still growing, Computer Vision remains in an infant stage, given that autonomous cars still can’t differentiate between a rock and paper bag or worse, a person and a static object. Despite this, there are plenty of startups on the scene working toward a solution.

As Artificial Intelligence increasingly blends with Biometric technology, it lends itself more easily to image recognition, allowing computers to correctly match fingerprints and facial patterns. But, no longer is it just matching two images. Now, it’s being taught to learn the difference between static images and liveness. This could prove invaluable for:

  • Spotting weapons
  • Suspicious behaviors
  • Dangerous object detection

Safety Begins at Home


With products such as security cameras prevailing within the security industry, businesses are hard at work creating and improving their products using the latest technologies. 

One such company is working to boost the clarity of their home security cameras. Think grainy gray, blurry images from night vision options or overly bright and distorted in the day. Their camera chip will be HDR and will be able to take much clearer pictures even in low-light. At home, you might find this product in a doorbell camera which could prove quite useful for smart homes which offer the option of allowing service people into your home from a remote app on your phone. 

And these cameras aren’t limited to your front door:

  • ATMs
  • Indoor/Outdoor cameras
  • CCTV Cameras
  • Number Plate Recognition Cameras

Though cameras have played a role in all of these areas for some time, the idea now is to keep them from being hacked and causing damage on both a product and a personal level. And like any type of Artificial Intelligence recognition system, these camera applications are created with advanced features to protect against hacking as well super speed processing of whether or not an object is an animal, a shadow, or some kind of inclement weather. 

The best part? We’re only a few months away from products which can boost the benefits of many of available security cameras. 

The Caveat of Image Recognition Systems

More and more people around the world own security cameras, not the least of which is security personnel. With this increased level of ownership, the market is expected to have over $20 billion in revenue in less than five years. With such high demand, it’s no wonder only about five percent of footage ever gets viewed. But what if security professionals could navigate up to 80 video sources on a single server? What if the algorithms, analytics, and video processing worked with cameras of all types including, but not limited to: 

  • Car dashboards
  • Body vests
  • Drone mounts

With all these cameras surrounding us, the one thing to watch is that it is still humans who input the information the cameras use to process. When it comes to security professionals such as law enforcement, investigation, and government authorities its import to make sure the camera doesn’t discriminate or operate on bias.

This an important issue even that even our most advanced intelligence agencies admits the algorithms for their facial recognition software are wrong about 15 percent of the time. However, one organization is working to improve and make these algorithms more ethically developed when it comes to image recognition as well as respecting individual privacy concerns. 

As we navigate the growing pains of new technologies, it's important to understand these solutions are meant to ensure customers, the public, and communities can trust the solutions being created.

If you’re looking for a new role in Computer Vision, take a look at our latest opportunities or get in touch with one of our expert consultants:

For our West Coast Team, call (415) 614 - 4999 or send an email to sanfraninfo@harnham.com.  

For our Mid-West and East Coast Teams, call (212) 796 - 6070 or send an email to newyorkinfo@harnham.com


Related blog & news

With over 10 years experience working solely in the Data & Analytics sector our consultants are able to offer detailed insights into the industry.

Visit our Blogs & News portal or check out our recent posts below.