The goal of machine vision is to enable computers to interpret and process visual information. The challenge is that the system has to extract valuable data from millions of numbers and make decisions based on that.
\begin{bmatrix} 80 & 7 & 28 & \dots & 35\\ 85 & 87 & 45 & \dots & 7\\ 96 & 66 & 40 & \dots & 83\\ 75 & 57 & 81 & \dots & 5 \\ \vdots & \vdots & \vdots & \ddots & \vdots \\ 35 & 26 & 27 & \dots & 74 \end{bmatrix}
What is happening in the picture?
What is the mood of the picture?
Which way is the other car going?
What is the weather like at the moment?
Character recognition (OCR)
Facial recognition and detection
Tracking
Visual effects (VFX)
Advanced driver-assistance systems (ADAS)
Image analysis (e.g. Healthcare)
Augmented reality (AR)
Security systems
Artificial intelligence as a concept is nowadays more of a marketing tool, it can be applied to almost anything, it is not possible to define its meaning precisely. Machine learning, on the other hand, is a concept that is part of AI but has a clear meaning. In image processing, a lane departure warning system can be created using classical methods, looking for nearly the same pattern in a given area of the camera image. A sign recognition program, on the other hand, requires training, since different signs can have different shapes, colours and sizes in the captured image, so the program has to learn to recognise these variations.
It is now easy to create machine vision systems that can perform a particular task faster and possibly with greater accuracy than humans. However, it is still a long way off before an artificially created system can be as versatile as human vision, being able to complete so many different tasks.