Article: Will Computers See Colors Differently than People?

Photo Source: Wikimedia Commons

With potentially infinitely more expansive vision, machines, with the help of artificial intelligence, may perceive the world in entirely different colors than people could imagine. We see differences in the way people describe color around the world. Different peoples have different views of the rainbow. Some have three words describing the colors of the rainbow, and most Western countries have seven colors designated with the acronym ROYGBIV: red, orange, yellow, green, blue, indigo, and violet. Color is an assignment of a word to describe the visual perception of a physical phenomenon known as an electromagnetic wave or light.

Scientists have determined color using instruments that can detect electromagnetic waves of varying frequencies. The visible spectrum that humans see ranges between 480 and 670 teracycles per second, a measure also known as Hertz. There are higher frequencies of electromagnetic radiation, such as ultraviolet, and very high-frequency waves, such as x-rays. At the other end of the spectrum, we find longer waves, such as infrared, which we can feel but not see, and even longer wavelengths, such as radio waves and microwaves.

Using the ocean as a metaphor, you can think of wavelength as the time between when a wave’s crest and its trough pass a fixed point like a buoy or it hits the shore. Larger ocean waves oscillate around every 14 seconds or 0.07 Hz, which is significantly slower than electromagnetic waves in the visible spectrum that range from 480 to 670 terahertz. One terahertz equals 1,000,000,000,000 oscillations or 1 trillion cycles per second.

People do not perceive all types of electromagnetic radiation. We see white light that separates into a rainbow of colors when it passes through a prism or droplets of water in the atmosphere. Our perception of electromagnetic radiation with the naked eye starts at red just above the infrared region and ends with violet just below the ultraviolet region. Some animals, such as insects and many birds, see ultraviolet light in addition to the visible spectrum. The birds we see do not look the same as how birds see each other. For example, in some bird species, the male and female birds may look the same in the visible range but wildly different in the ultraviolet range. When we see a pigeon in flight, the wings do not look terribly remarkable, but adding in the ultraviolet range of light that birds see, the wings have additional patterns and spots. 

Our color vision stems from the physical nature of our light-sensing retina in our eyes. Our retina has remarkable cells called rods and cones due to their shape when viewed under a microscope. The more sensitive rods detect light across the spectrum without differentiating color. The rods help us see in low light, which we see in black and white when it is dark. However, the less sensitive cones separate into three groups—blue, green, and red--based on the frequency of light they detect. Different wavelengths of light stimulate one or more cones to produce the perception of color. For example, we perceive yellow when the red and green cones get stimulated simultaneously. All the cones at once produce the perception of white. Now, birds and insects have a fourth cone for ultraviolet light, meaning they perceive a variety of different “colors” that we cannot see. We can infer these colors but cannot truly see them as birds do.

The nature of vision and its limits based on the types of detector cells in animal eyes implies that the visible world for humans has limitations compared to birds and insects based on the capabilities of our rod and cone cells. Because people can make detectors that span the entire electromagnetic spectrum from radio waves through visible light to x-rays, machines with full spectrum vision would “see” infinitely more “colors” than people could ever imagine. Currently, computer vision depends on training the computer to assign colors as people see them. Still, the potential for a computer to see the world in a wider light, even wider than birds and insects, creates a potentially widening gulf between human perception and computer perception in the spectrum of electromagnetic radiation. Would computers then see a larger reality than people, and how would that affect the machine’s interpretation of the world and beyond? It could separate human and computer interpretation of events and actions in an unpredictable way.

Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high-level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.

You can buy his book on Amazon in paperback and in kindle format here.