A Pulse-Coupled Neural Network as A Simplified Bottom-Up Visual Attention
Model
Marcos Quiles, Roseli Romero, and Liang Zhao
Department of Computer Science
Institute of Mathematics and Computer Science
University of S˜ ao Paulo - S˜ ao Carlos, SP, Brazil
{ quiles, rafrance, zhao }@icmc.usp.br.
Abstract
This work presents a bottom-up visual attention model
based on a Pulse-Coupled Neural Network for scene seg-
mentation. Each object in a given scene is represented by a
synchronized pulse train, while different objects fire at dif-
ferent phases. Taking this into account, the model focuses
on one object at a time. Using this model, the limit of lin-
ear non-separability can be easily overcome and computer
simulations show its good performance.
1 Introduction
The development of computer vision systems still probe
to be a difficult task for computer scientists due to all the in-
volved steps, as well as the complexity and great amount of
data that needs to be analysed at the same time. To reduce
the amount of incoming visual data obtained from the envi-
ronment, the human visual system is able to select only the
relevant amount of information necessary to perform tasks
such as object recognition [16]. Visual Attention is the way
that humans and other biological systems use to select rele-
vant visual information [15].
Selective visual attention is used by the biological sys-
tem to optimize its processing capacity. Selective attention
is responsible for identifying the part of the visual input
where the processing is performed and, at the same time
suppressing irrelevant visual information [2]. Selective vi-
sual attention is generated by a combination of information
from the retina and early visual cortical areas (bottom-up
attention) as well as feedback signals from areas outside of
the visual cortex (top-down attention) [8].
The bottom-up attention is formed by simple features ex-
tracted from the image, such as intensity, stereo disparity,
colour, orientation, and others. All this information is com-
bined to create a saliency map which represents the points of
interest in the visual input. The top-down attention signals
are responsible for being biased concerning the competition
of all the points generated by the saliency map. This infor-
mation can be, for example, a visual search for a specific
object.
Because of this biological evidence, selective visual at-
tention can be used to develop artificial vision systems [2].
In this case, selective visual attention is used to reduce the
amount of incoming data by selecting only part of the vi-
sual information for further processing improving the per-
formance or efficiency of the system.
Several models of visual attention have been proposed
and they can be divided in two groups. The first belongs to
computational neuroscience area, where the computational
models are used to simulate and understand biological sys-
tems [5, 4, 8].
The second group is related to computer vision area,
where the models are applied to reduce the amount of data
analysed by the system [16, 10]. In this case, the models do
not need to have a strong biological plausibility.
The model proposed in this paper belongs to the sec-
ond group. Two aspects motivate the present work. The
first is to explore the parallel architecture typically observed
in neural network models and the second is to use the de-
velop model to face nonlinear separable problems, such as
the Double Spiral problem [3].
In this paper, a visual attention model based on a Pulse-
Coupled Neural Network is proposed. The Pulse-Coupled
Neural Networks (PCNN) is a neural network composed of
spiking neurons. It is based on the Linking-field neural
network model proposed by Eckhorn [6]. The Linking-
field neural network was projected as a minimal model to
explain the experimentally observed synchronous feature-
dependent activity of neural assemblies over large cortical
distances in the cat cortex [9]. Considering that this neural
network was proposed to study the cat visual cortex, that is
the part of the brain that processes the information deriving
Proceedings of the Ninth Brazilian Symposium on Neural Networks (SBRN'06)
0-7695-2680-2/06 $20.00 © 2006