[Visual recognition with early and delayed observer tuning to a specific set of images]

It has been shown in psychophysiological experiments that re-tuning of the observer's visual system from a long set to a short one improves recognition to a greater extent after an anticipating signal about the set than after a delayed one. The change of the anticipating signal to a delayed one increases the number of "I don't know" responses, while the number of errors of confusing resembling figures remains unchanged. Temporal limitations of processing of the information about the stimulus increases the number of confusion errors. The degree of the observer's training reflects upon the number of errors; the number of "I don't know" responses does not change. The facts obtained attest that two processes are involved in recognition: one of them is sensitive to the position in time of the signal about the set and does not depend on training. The other process, on the other hand, does not depend on the signal position in time and changes depending on training. It is assumed that the first process is a complete element-to-element description, while the second one is a classification process.