In this paper, a novel model of object-based visual attention extending Duncan's Integrated Competition Hypothesis [24] is presented. In contrast to the attention mechanisms used in most previous machine vision systems which drive attention based on the spatial location hypothesis, the mechanisms which direct visual attention in our system are object-driven as well as feature-driven. The competition to gain visual attention occurs not only within an object but also between objects. For this purpose, two new mechanisms in the proposed model are described and analyzed in detail. The first mechanism computes the visual salience of objects and groupings; the second one implements the hierarchical selectivity of attentional shifts. The results of the new approach on synthetic and natural images are reported.

this article. We also thank Stephan Dekker for technical assistance

Working memory is a system that keeps limited information on-line for immediate access by cognitive processes. This type of active maintenance is important for everyday life activities. The present study shows that maintaining a location in spatial working memory affects the trajectories of saccadic eye movements toward visual targets, as the eyes deviate away from the remembered location. This finding provides direct evidence for a strong overlap between spatial working memory and the eye movement system. We argue that curvature is the result of the need to inhibit memory-based eye movement activity in the superior colliculus, in order to allow an accurate saccade to the visual target. Whereas previous research has shown that the eyes may deviate away from visually presented stimuli that need to be ignored, we show that the eyes also curve away from remembered stimuli.

Alerting, orienting, and executive control are widely thought to be relatively independent aspects of attention that are linked to separable brain regions. However, neuroimaging studies have yet to examine evidence for the anatomical separability of these three aspects of attention in the same subjects performing the same task. The attention network test (ANT) examines the effects of cues and targets within a single reaction time task to provide a means of exploring the efficiency of the alerting, orienting, and executive control networks involved in attention. It also provides an opportunity to examine the brain activity of these three networks as they operate in a single integrated task. We used event-related functional magnetic resonance imaging (fMRI) to explore the brain areas involved in the three attention systems targeted by the ANT. The alerting contrast showed strong thalamic involvement and activation of anterior and posterior cortical sites. As expected, the orienting contrast activated parietal sites and frontal eye fields. The executive control network contrast showed activation of the anterior cingulate along with several other brain areas. With some exceptions, activation patterns of these three networks within this single task are consistent with previous fMRI studies that have been studied in separate tasks. Overall, the fMRI results suggest that the functional contrasts within this single task differentially activate three separable anatomical networks related to the components of attention.

Even though it is undisputed that prior information regarding the location of a target affects visual selection, the issue of whether information regarding nonspatial features, such as color and shape, has similar effects has been a matter of debate since the early 1980s. In the study described in this article, measures derived from signal detection theory were used to show that perceptual sensitivity is affected by a top-down set for spatial information but not by a top-down set for nonspatial information. This indicates that knowing where the target singleton is affects perceptual selectivity but that knowing what it is does not help selectivity. Furthermore, perceptual sensitivity can be enhanced by nonspatial features, but only through a process related to bottom-up priming. These findings have important implications for models of visual selection.

The tension between privacy and awareness has been a persistent difficulty in distributed environments that support opportunistic and informal interaction. For example, many awareness systems that display ‘always-on ’ video links or PC screen contents have been perceived as too invasive, even though functional real-world analogues, like open-plan offices, may provide even less privacy than their online counterparts. In this paper we explore the notion of privacy in open-plan real-world environments, in order to learn more about how it might be supported in distributed systems. From interviews and observations in four open-plan offices, we found that attention plays an important role in the management of both confidentiality and solitude. The public nature of paying attention allows people to build understandings of what objects in a space are legitimate targets for attention and allows people to advertise their interest in interaction. Our results add to what is known about how privacy works in real-world spaces, and suggest valuable design ideas that can help improve support for natural privacy control and interaction in distributed awareness systems.

In functional neuroimaging studies of episodic retrieval (ER), activations in prefrontal, parietal, anterior cingulate, and thalamic regions are typically attributed to episodic retrieval processes. However, these activations are also frequent during visual attention (VA) tasks, suggesting that their role in ER may reflect attentional rather than mnemonic processes. To investigate this possibility, we directly compared brain activity during ER and VA tasks using event-related fMRI. The ER task was a word recognition test with a retrieval mode component, and the VA task was a target detection task with a sustained attention component. The study yielded three main findings. First, a common fronto-parietal-cingulate-thalamic network was found for ER and VA, suggesting that the involvement of these regions during ER reflects general attentional processes. This idea is compatible with some of the interpretations proposed in the ER literature (e.g. postretrieval monitoring), which may be rephrased in terms of attentional processes. Second, several subregions were differentially involved in ER versus VA. For example, the frontopolar cortex and the precuneus were more activated for ER than for VA, possibly reflecting retrieval mode and processing of internally generated stimuli, respectively. Finally, the study yielded an unexpected finding: some medial temporal lobe regions were similarly activated for ER and VA. This finding suggests that the medial temporal lobes may be involved in indexing representations within the focus of consciousness, regardless of whether they are mnemonic or perceptual. Overall, the present results suggest that many of the activations attributed to specific cognitive processes, such as episodic memory, may actually reflect more general cognitive oper...

& Attention operates at an early stage in some experimental paradigms and at a late stage in others, which suggests that the locus of selection is flexible. The present study was designed to determine whether the locus of selection can vary flexibly within a single experimental paradigm as a function of relatively modest variations in stimulus and task parameters. In the first experiment, a new method for assessing the locus of selection was developed. Specifically, attention can influence perceptual encoding only if it is directed to the target before a perceptual representation of the target has been formed, whereas attention can influence postperceptual processes even if attention is cued after perception is complete. Eventrelated potentials were used to confirm the validity of this method. The subsequent experiments used cueing tasks in which subjects were required to perceive and remember a set of objects, and the difficulty of the perception and memory components of the task were varied. When the task overloaded perception but not working memory, attention influenced the formation of perceptual representations but not the storage of these representations in memory; when the task overloaded working memory but not perception, attention influenced the transfer of perceptual representations into memory but not the formation of the perceptual representations. Thus, attention operates to select relevant information at whatever stage or stages of processing are overloaded by a particular stimulus– task combination. &

On behalf of:

Human vision uses mechanisms of covert attention to selectively process interesting information and overt eye movements to extend this selectivity ability. Thus, visual tasks can be effectively dealt with by limited processing resources. Modelling visual attention for machine vision systems is not only critical but also challenging. In the machine vision literature there have been many conventional attention models developed but they are all space-based only and cannot perform object-based selection. In consequence, they fail to work in real-world visual environments due to the intrinsic limitations of the space-based attention theory upon which these models are built. The aim of the work presented in this thesis is to provide a novel human-like visual selection framework based on the object-based attention theory recently being developed in psychophysics. The proposed solution -- a Hierarchical Object-based Attention Framework (HOAF) based on grouping competition, consists of two closely-coupled visual selection models of (1) hierarchical object-based visual (covert) attention and (2) object-based attention-driven (overt) saccadic eye movements. The Hierarchical Object-based Attention Model (HOAM) is the primary selection mechanism and the Object-based Attention-Driven Saccading model (OADS) has a supporting role, both of which are combined in the integrated visual selection framework HOAF.