Video transcript
Seeing and generating images
Welcome to SINet, headquarters of the high-tech spy agency, Stratum Intelligence Network. SINet uses retinal
scanners at all entrances to identify agents and prevent unauthorized entry.
In addition to the retinal scanners at the entrances, SINet uses other computer vision AI tools to keep their
headquarters secure. Computer vision AI tools use three main skills to function: recognizing shapes, classifying
objects, and generating new images. Let’s follow Agent Hawk as he attempts to enter the building.
As Agent Hawk positions his face close to the retinal scanner, the computer vision AI scans his eye. Computer
vision AI doesn’t “see” an eye; it interprets pixels. The system recognizes shapes by analyzing features such as
circles, curves, edges, and lines. This analysis is the foundation for interpreting the data. The system detects the
roundness of the pupil and unique patterns in the iris, much like noticing the shape of a familiar object.
Next, the system compares the shape of Agent Hawks’ eye to a secure database of enrolled agents. The pattern
scanned by the computer vision AI matches Agent Hawk’s profile, so the system labels the scan as authorized and
grants Agent Hawk entry access to the building. If the scan had been labeled as unauthorized, the system would
have raised an alert for review and denied entry. This process is known as object classification, which involves
categorizing objects accurately based on their distinctive features.
Now, let’s follow Agent Hawk to the spy lab. In the lab, technicians are using computer vision AI to generate new,
realistic eye pattern images by mixing features learned from many scans. The computer vision AI is generating new
images based on what it has already learned. This approach enables the agency to test the security system using
simulated scans and improve the system’s reliability and security.
Each time the retinal scanner works through the process of recognizing shapes during a scan, classifying objects
such as the roundness of a pupil and the unique features of an iris, or generating new images for testing, computer
vision AI helps keep the agency secure and up to date. This same computer vision AI approach is used daily, from
unlocking smartphones to routing packages or from reading medical images to generating art. Computer vision AI
recognizes shapes, classifies objects, and generates images, enhancing your ability to improve and innovate.
Page 1 of 1