A computational model of perceptual reversal alternating between two interpretations is presented. Initially, the model represents the ambiguous state of a reversible picture, such as the bistable face-vase image. The internal state of the network evolves to settle into a stable state, which corresponds to one of two alternatives. Top-down feedback proves a deciding factor in leading the system into a modeled perceptual state over the time course. At any given time, top-down input from temporal associative memory provides contextual modulation of bottom-up input in the network. The model accounts for the role of top-down knowledge in resolving perceptual ambiguity as well as reversibility from one state to another.
Techniques to identify an attentive object in an image, captured synchronously with an eye tracker, rely on the human fixations being inside the boundary of the object. However, eye tracking experiments show that fixations are often near, but outside an attentive object, when the viewer is not trying to place the fixations inside. If fixations occur outside an object boundary, it is difficult to segment it properly. We address this problem by correlating context-directed visual saliencies with eye fixations. Our model is based on prior specification of the weights applied to bottom-up saliency. Experiments on the meeting scenes indicate that our method achieves a detection rate of 95.6%.
A common assumption in visual attention is based on the rationale of "limited capacity of information processing". From this view point there is little consideration of how different information channels or modules are cooperating because cells in processing stages are forced to compete for the limited resource. To examine the mechanism behind the cooperative behavior of information channels, a computational model of selective attention is implemented based on two hypotheses. Unlike the traditional view of visual attention, the cooperative behavior is assumed to be a dynamic integration process between the bottom-up and top-down information. Furthermore, top-down information is assumed to provide a contextual cue during selection process and to guide the attentional allocation among many bottom-up candidates. The result from a series of simulation with still and video images showed some interesting properties that could not be explained by the competitive aspect of selective attention alone.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.