I'm thinking about trying to implement object recognition/tracking using a silicon retina (also known as the Dynamic Vision Sensor) and using the Semantic Pointer Architecture as my cognitive model. However, I want to make sure that it is biologically plausible, that is it respects the experimental limitations observed with human subjects.
Consequently, what human limitations on object tracking and recognition have been observed?
I think you will want to look into ecological psychological models for perception.
Some cornerstones include 'the gorila study' show inattentional blindness
Others are among the many sub-phenomenon with Persistence of Vision studies, many of which now focus on the a concept called Iconic Memory. These matters deal with misapprehension of (usually incomplete) visual stimuli due to the action of cognition to create a meaningful recogition (e.g. probably what is 'ready for encoding / memory'). This also generally covers the phenomenology where we receive too-few stilimuli (objects moving too fast), and must cognize the available information to create meaning.
Here are a few good perspectives on human vision. This is far from being a comprehensive set of human vision limitations, but if you dig into their references, you will find much of the community that studies this topic.
Shaffer, D., Dolgov, I., Maynor, A., & Kelly, K. (in press). Blind(fold)ed by science: A constant target-heading angle is used in visual and nonvisual pursuit. Psychological Science.
Schwark, J., Sandry, J., McDonald, J., & Dolgov, I. (2012). False feedback in vigilance tasks. Attention, Perception, & Psychophysics, 74, 1583-1589
Dolgov, I., McBeath, M., & Sugar, T. (2009). Evidence for axis-aligned motion bias: Football axis-trajectory misalignment causes systematic error in projected final destinations of thrown American footballs. Perception, 38(3), 399-410.
(source / disclosure, I know some of the authors of the above papers)
