Computer Vision


Computer vision starts thinking by focusing, just like people do. First, it scans millions of data, focuses on searches matched with defined criteria, then attributes meaning to things. Finally, it becomes able to differentiate a cat from a dog. It goes even further. Computer vision that is able to distinguish our emotions, can also differentiate our facial expressions as happy, unhappy, nervous or confused 1. Measuring the length of queues 2, or finding the missing 3 are ordinary jobs for computer vision. There are also many applications in the field of medical science 4. Most of the time, adjustments are needed to achieve clear images 5.

1 Monitoring Emotions

Tracking and observing capabilities of computer vision can evaluate emotions. The pilot project of Chinese government has started in Hangzou last year. The application will assess primary school children’s concentration in class from their facial expressions to make learning more efficient.


2 Queue Management

Customer satisfaction is a must for the service sector. It does not matter whether it is a cashier queue or a ticket line at the airport. The first thing we cannot tolerate in the age of speed, is waiting. Therefore, computer vision comes into play. Counting the number of people in the queue and managing wait duration become important to tools in increasing customer satisfaction.


3 Search & Rescue

Thanks to thermal cameras that can detect body heat, a missing person wandering into a cornfield was found without injury.

Public safety cameras are a sine qua non in large open spaces. If there is a missing child reported to security, one of the first things to be done is to enter detailed information about the child in the system and ensure computer vision focus. It is important to be able to identify children with red coats and 130 cm height when there are hundreds of people moving or standing in the street. Our brains’ constant unconscious focusing process in everyday life, is actually the first step in solving complex structures.

4 The Eye in Me

Microsoft is using machine learning and computer vision so radiologists can get a more detailed understanding of how a patient's tumor is progressing.

As imaging techniques progress in healthcare, the amount of data generated yearly increases exponentially. One of the biggest problems at the moment is the limited number of medical staff who review and analyze these images and the need for long years of serious specialization. Microsoft’s InnerEye project is based on a system that examines these three-dimensional radiology images with computer vision. On average, working 40 times faster than a human being, this system focuses on the exact problem and analyzes the images according to various criteria.


5 Image Adjustment

One of the most distinctive features of computer vision is its ability to disregard image distortions or normally unexpected details. For example, it will be a critical step for a camera to recognize hail or a change in lighting due to snow and filter out these details. This is how Waymo, specialized in autonomous vehicles, eliminates factors that deteriorate camera images such as raindrops or snowfall.