A Neural Model of Attention and Feedback for Computing Perceived Brightness in Vision

Document Type

Book Chapter

Publication Title

Handbook of Neural Computation


This chapter describes the Attentive Vision Filter (AVF) model, in which attention and feedback play an important role in modelling brightness perception. This model has the capacity to explain a wide variety of brightness illusions which include both contrast and assimilation of brightness, while at the same time having a strong neural correlate in the parallel visual channels. This chapter also demonstrates, via scale considerations, why a non-linear model is necessary to explain the phenomenon of brightness perception. This chapter concludes by noting how both of these, i.e. attention and scaling, may be achieved in the brain through feedback signals from the cortex to the thalamus. This observation in turn could be implemented in Computer Vision by introducing a computational feedback mechanism into the traditional Deep Convolutional Neural Networks.

First Page


Last Page




Publication Date


This document is currently not available here.