Naslov
Lightness anchoring in neural networks

Autori
Domijan, Dražen

Izvornik
Review of Psychology (1330-6812) 8 (2001), 1-2; 3-11

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
Lightness perception; Anchoring; Neural network; Presynaptic inhibition

Sažetak
Lightness anchoring is a process that transforms the scale of relative luminance ratios into a scale of perceived shades of greys. Recent psychophysical investigations suggest that a surface with the highest luminance in a scene appears white. In order to provide a mechanistic account of this process, a new neural network is proposed based on presynaptic inhibition of feedforward inhibitory pathways and self-excitatory feedback. Presynaptic inhibition serves as a gate that controls the amount of inhibition that cells may receive. Computer simulations showed that the proposed model correctly simulates human psychophysical data on the staircase Gelb effect, and the role of field size and insulation on its appearance. However, the model is not able to account for the perception of luminosity and the role of articulation in lightness perception. A neural network for size estimation is presented which may partially circumvent the problem, but how such a network interacts with an anchoring network in order to provide complete representation of lightness values is not yet clear.

Izvorni jezik
Engleski

Znanstvena područja
Psihologija

POVEZANOST RADA