Vision & AI

Computer Vision is the science and technology of obtaining models, meaning and control information from visual data.  The two main fields of computer vision are computational vision and machine vision.  Computational vision has to do with simply recording and analyzing the visual perception, and trying to understand it.  Machine vision has to do with using what is found from computational vision and applying it to benefit people, animals, environment, etc.

Computer Vision has influenced the field of Artificial Intelligence greatly.  The Robocup tournament and ASIMO are examples of Artificial Intelligence using Computer Vision to its greatest extent. The Robocup tournament is a tournament for robot dogs playing soccer.  To be able to play soccer, these dogs must be able to see the ball, then react to it accordingly.  Engineers of these robot dogs have been challenged to create robot dogs who can beat the best soccer players at soccer in around fifty years.

ASIMO, seen below, is another example of how computer vision is an important part of Artificial Intelligence.  ASIMO is a robot created by Honda, but of course, all robots need to be able to know where to move around and what is in its surroundings.  To be able to do this, ASIMO uses cameras to visualize computationally what is in its surroundings, then uses it to achieve its goal.

Honda's ASIMO playing with a ball.ASIMO walking down stairs.

Artificial Intelligence can also use computer vision to communicate with humans.  GRACE the robot, shown below, is a robot who could communicate slightly with humans to be able to recognize her surroundings and achieve a specific goal.  For example, GRACE attended a conference through a lobby and up an elevator by communicating with humans.  Communications included understanding that she had to wait in line, and asking others to press the elevator button for her. She also has a binocular vision system allowing her to react to human gestures as well.

The robot named GRACE.

Artificial Intelligence also uses computer vision to recognize handwriting text and drawings.  Text typed down on a document can be read by the computer easily, but handwritten text cannot. Computer vision fixes this by converting handwritten figures into figures that can be used by a computer.  An example is shown below.  The attempted drawing of a rectangular prism resting on three other rectangular prism is converted by computer vision to a 3-D picture of the same thing, but in a format usable by the computer and more readable by users.

Handwritten cubes converted into usable figures.

Another important part of Artificial Intelligence is passive observation and analysis.  Passive observation and analysis is using computer vision to observe and analyze certain objects over time. For example, in the pictures below, on the first one, the passing cars are being observed and analyzed as what type of car by the computer.  This can be done by outlining the car shape and recording it.  On the second picture, the flock of geese are observed and analyzed over time.  The record could serve to predict when geese would come again, for how long they would stay, and how many of them there could be.

Cars being analyzed by their shape.A flock of geese observed and analyzed by computer vision.