The other aspect of subjectivity is that color (as in color perception) is dependent upon context, both in space and time.
The retina of tetrapods performs the most sophisticated neural processing that exists outside of the brain. A key aspect of this processing is differencing -- both spatial and temporal differencing. We can infer a couple of reasons for differencing: extraction of key perceptual cues and data reduction.
Spatial differencing is key for edge detection. In machine vision, edge detection is a primitive function upon which higher level functions are built. This seems to be the case with the human visual system as well. Recognition of objects, for example, relies on edge detection.
Spatial differencing is built right into the back of the eye, where all the relevant information is readily available.
Temporal differencing is key to motion detection. Moving edges are of particular interest to most tetrapods, as moving objects may represent immanent threat or opportunity. Again, the basics of spatio-temporal differencing that enables motion detection occur in neurons at the back of the eye.
Another reason we can infer for this differencing is data reduction. The eye captures far too much information to be conveyed by the optic nerve. Extracting "what is different from before" represents a massive amount of data reduction.
These retinal primitives play a role in optical illusions. The name "retinal illusions" would be more accurate. Perceived brightness and saturation of a particular swatch can be manipulated by what is placed near that swatch. The eye isn't so much perceiving the swatch as "what about this swatch is different from its surroundings."
Similarly, persistent images of complementary color are an artifact of temporal differencing. "What about this is different from what was perceived a short time ago."
The eyes: lying to us since at least the Devonian.
|
The retina of tetrapods performs the most sophisticated neural processing that exists outside of the brain. A key aspect of this processing is differencing -- both spatial and temporal differencing. We can infer a couple of reasons for differencing: extraction of key perceptual cues and data reduction.
Spatial differencing is key for edge detection. In machine vision, edge detection is a primitive function upon which higher level functions are built. This seems to be the case with the human visual system as well. Recognition of objects, for example, relies on edge detection.
Spatial differencing is built right into the back of the eye, where all the relevant information is readily available.
Temporal differencing is key to motion detection. Moving edges are of particular interest to most tetrapods, as moving objects may represent immanent threat or opportunity. Again, the basics of spatio-temporal differencing that enables motion detection occur in neurons at the back of the eye.
Another reason we can infer for this differencing is data reduction. The eye captures far too much information to be conveyed by the optic nerve. Extracting "what is different from before" represents a massive amount of data reduction.
These retinal primitives play a role in optical illusions. The name "retinal illusions" would be more accurate. Perceived brightness and saturation of a particular swatch can be manipulated by what is placed near that swatch. The eye isn't so much perceiving the swatch as "what about this swatch is different from its surroundings."
Similarly, persistent images of complementary color are an artifact of temporal differencing. "What about this is different from what was perceived a short time ago."
The eyes: lying to us since at least the Devonian.
- mark 2-15-2015 7:51 pm