With a reasonable amount of teaching and practice, humans can read and comprehend text fairly easily. However, when it comes to vision we only perceive a small amount of what appears in our field of view. We can recognize the objects and details we focus on, but much of what is contained in our peripheries we either don’t recognize or we infer based on our surroundings. In addition, we can read at whatever pace we choose, but an image appears in front of us only for an instant with so much information to take in. Because we know the alphabet and understand phonetics we can read any text placed in front of us. We also have the aid of a dictionary to help us comprehend words which are unfamiliar.
It is fairly easy to program a computer to parse text. There is a relatively small amount of building blocks, including the alphabet, punctuation, and line breaks, which combine to make up any string. Images are more complicated. Using the technologies available we can break an image down into pixels and analyze the color characteristics and determine the position of objects in three dimensions. The sheer amount of data is enormous compared to a text file. What’s even more complicated is creating a program that can make sense of that data with precision. In the realm of text we can easily teach the computer to recognize the word “boat” with 100% accuracy, but writing a program that can recognize the image of a boat is difficult. Can we create a “visual dictionary” to aid with this? I think we’re on our way.
As we build up our visual vocabulary we will be able to achieve a higher awareness of our surroundings. Surveillance cameras connected to computers will be able to detect and report specific threats without human monitoring. Facial recognition will allow us to identify and obtain information about individuals we pass in the street. The ability to collect precise spatial data means we can use that data to re-create three dimensional spaces elsewhere. I would imagine there to be significant impact in the medical field as well. Smarter cameras will allow for better imaging, aiding doctors in diagnosis and in surgery.