Computational Models of Visual Attention: A Survey

Computational Models of Visual Attention: A Survey

Rajarshi Pal (Institute for Development and Research in Banking Technology, India)
DOI: 10.4018/978-1-4666-4558-5.ch004
OnDemand PDF Download:
No Current Special Offers


Even the enormous processing capacity of the human brain is not enough to handle all the visual sensory information that falls upon the retina. Still human beings can efficiently respond to the external stimuli. Selective attention plays an important role here. It helps to select only the pertinent portions of the scene being viewed for further processing at the deeper brain. Computational modeling of this neuro-psychological phenomenon has the potential to enrich many computer vision tasks. Enormous amounts of research involving psychovisual experiments and computational models of attention have been and are being carried out all within the past few decades. This article compiles a good volume of these research efforts. It also discusses various aspects related to computational modeling of attention–such as, choice of features, evaluation of these models, and so forth.
Chapter Preview


What is Visual Attention?

We, the human beings, are amazingly efficient in real-time interaction with the dynamic environment surrounding us. We, constantly, gather information about our surroundings through five senses. After analyzing or interpreting the information, the brain decides the course of action, but do we process all the incoming sensory information at the deeper levels of the brain? Let this question be put in a slightly different manner. All the time, our senses are actively sensing the outside world. All of that sensory information is routed towards the brain though nerves. But does all this information reach the deeper level of the brain where recognition and decision making takes place? For example, we cannot effectively listen to the important discussion on a television channel while we simultaneously carry out the conversation with a friend over the telephone. So, there is a limit to the number of things the brain can process simultaneously.

The human brain intelligently filters out the majority of incoming sensory information before it can reach the deeper levels of the brain. This phenomenon is known as attention. Attention helps us to attend only a selective subset of sensory information. In the pretext of the subject matter of this article, we limit the discussion only to visual attention. Even, only a few selected portions of the visual stimuli sensed in the retina of the eye are able to draw our attention.

Why is a Computational Model Required?

Like humans, computer vision tasks also face the challenge of dealing with huge amount of input (Tsotsos, 1990). The attention mechanism of human vision has influenced computer vision researchers to restrict the computation in certain areas of input. As a result, modeling visual attention draws significant research effort within the past few decades. It gathers theories from psychology, neurobiology of human visual system, and other related topics. Psycho-visual experiments have provided some theoretical reasoning for saliency of a location or an object. Computer vision researchers try to fit various types of models on acquired salient data on the basis of these psycho-visual experiments.

Selective attention to relevant salient locations in a scene has various advantages. It reduces the computational burden by decreasing the amount of data to be processed. Moreover, suppression of irrelevant information ensures influence of only the relevant locations of the scene in the outcome of the system.

Complete Chapter List

Search this Book: