The Gollin incomplete figure test as a masking problem

The Gollin test of incomplete figure perception is usually employed to measure the thresholds of recognition in children and adults, and to study a process which provides a basis for the perception of incomplete figures as Gestalts (Foreman and Hemmings, 1987 Perception 16 543 ^ 548). Here we suggest that this test, along with such tests as the Poppelreuter test of figure extraction, and the Mooney faces test, may be considered as a visual masking problem. Digital image processing allows us to measure the spatial properties and spatial-frequency spectrum of the absent part of the image as a mask. We compare incomplete masking with other traditional types of masking. Using a noise paradigm, we have measured the signal-to-noise ratio for incomplete figure perception in normal participants and in neurological patients. This is the most powerful aspect of this new approach. Clinically, the new paradigm may provide a quantitative measure of agnosia. We have developed the hypothesis that some forms of visual agnosia arise primarily from an especially high level of noise within higher visual processing, including memory systems. We classify this type of agnosia as ‘filtration agnosia’. The concept of incomplete figure perception as noise filtration is therefore important for clinical purposes.

European Conference on Visual Perception.
Glasgow, UK.
25-29 August 2002.