There are plenty of computer vision examples out there which demonstrate that computers can now identify and label images. Some of the most interesting computer vision use cases are in the following industries:
- Retail and security
In this article, we will discuss some interesting uses of computer vision in order to provide you with an understanding with some practical applications on the market today and to help you determine if AI could be a solution for your business needs.
Computer Vision for Retail Stores
Not too long ago, Amazon unveiled a store, where shoppers can bypass the lines and pay for their merchandise right away. Computer vision facial recognition cameras are used to let employees know when something was taken from the shelves, whether or not it was returned and remove an item from the shopping cart if it was returned. Also, the cameras track each person inside the store all times making sure that each shopper pays for the merchandise they selected and verifying that all billing transactions are correct. After the shopper has added everything into their virtual basket, they can simply leave the store and they will be billed by Amazon later on.
In terms of computer vision in retail fashion, Amazon has developed a virtual mirror that uses a blended reality display that places a picture or the customer into a virtual scene and into virtual clothing. This technology uses facial detection to track the user’s eye motion since it uses this information to determine what the user is looking at.
Automotive Computer Vision
Recent data by the World Health Organization shows that more than one million people are killed every year in car accidents and this such a pace is not expected to slow down anytime some and is expected to continue into the next decade and beyond. One of the main factors in car accidents is just plain old human error and lack of attention. Computer vision in the automotive industry is being used to do something about this problem. And one of the most interesting use cases is a company called Waymo. If you have been following computer vision in self-driving cars that made headlines recently, you probably heard of Waymo. Basically, it is trying to optimize transportation by going beyond cars that drive themselves and equipping them with cameras that can spot pedestrians ensures that it can detect movement of pedestrians from 300 yards away. Also, it trained the computer vision with by driving seven million miles on public roads so they can safely navigate in everyday traffic.
For example, let’s say that you are driving along and a cyclist in front of you extend his/her hand indicating that they would like to shift over to your lane. The software will detect the hand motions and make a determination if in fact this a request to shift over to your lane, or merely a random motion such as a stretch. If it evens it as a request to shift lanes, it will instruct the car to slow down and let the cyclist through. It uses deep networks in order to predict plan, map out and simulate various situations in order to train the vehicles how it should act in various situations.
Computer Vision in Healthcare
Computer vision is used to identify and diagnose conditions and illnesses and make lifesaving medical interventions. There have been some arguments in healthcare which one is better: computer vision vs sensors for smart healthcare. There is really no need to pin one against the other because computer vision must be used along with the sensors to producing better results. For example, Gauss Surgical has developed a solution that monitors blood loss in real time. The sensors detect the amount of blood located on surgical sponges which are then processed by machine learning algorithms which make a determination how much blood was lost. The technology is currently being used in surfing surgical operations and Caesarian deliveries.
Industrial Uses of Computer Vision
Industrial computer vision software development is used to keep an eye on the state of industrial sites such as factories, remote wells, and any other strategic sites. An interesting use case was implemented by Osprey Informatics, which employed computer vision to monitor remote oil wells in order to eliminate unnecessary visits to a well site. The computer vision system provides images of the customer’s site 15-minute apart, with added options for live videos and real-time images. As a result, site visit costs were cut in half and the average cost for an employee to make a visit now costs 20 times less.
Computer Vision in Agriculture
Believe it or not, the agriculture industry is also employing computer vision technology to make operations more efficient such as growing methods, yielding more crops and generating higher profits. An interesting use case is a company called SlantRange, which uses drones with computer vision cameras to scan the crops and determine whether or not they are under threat. The drone hovers at an altitude of about 400 feet with a 4.8 cm/pixel resolution camera. Once it is airborne, the camera takes pictures of the crops which help identify possible hazardous conditions such as infestation, lack of water nutrition. It also makes estimates what the crop yields will be when it is time to harvest. All this data is funneled into an analytical system which provides data insights and allows farmers to take action in order to save their crops.
Computer Vision Applications in Banking
The banking has used image recognition software to prevent fraud by authenticating documents via machine learning. For example, let’s say that customer wants to deposit a check but does not feel like making a trip to the bank. All they have to do is take a picture of the check with their phone or tablet and the bank’s computer vision software analyzes its authenticity. As soon as the system has verified the check, it is deposited to the customer’s account.
While some industries have been ahead of others in terms of computer vision implementation, we are seeing greater adoption across the board. However, everything that is available today still relies on humans to supervise, analyze, offer insights and make a decision and take action. Even though certain companies are developing technology that would allow for self-driving cars, the reported fatality rates demonstrate that it still some time away from being commercially available. We are seeing the same in other industries as well where computer science is playing a big role, but the human factor is still necessary as well. As this technology continues to evolve and research learn to fine tune it, soon or later, it will replace humans in pretty much every aspect.
Check my BlockDelta profile for additional articles.