| Summary: | 碩士 === 國立成功大學 === 醫學工程研究所碩博士班 === 90 === The influence of visual attention on human visual experiences is significant for cognitive psychology. In this thesis, visual attention is denoted by selective attention, defined as follows: the allocation of limited processing resources to some stimuli of tasks at the expense of others. Visual attention deficits are usually accompanied with increasing response time and decreasing correct rate.
The study is motivated by the following observations: (1) In terms of cognitive rehabilitation, practitioners often need to predict attention response in real world scenes for strategies of clinical interventions. In studies on visual attention, a program of research, in which subjects are asked to visual search for the specific target, has been conducted for capturing responses in specific visual stimulus. The result, a profile, is compared with a standard profile for estimating visual attention. However, this standard profile usually is based on the statistical analysis, which is a costly and time-consuming process. (2) Although these studies have provided much valuable information on various patterns between visual stimulus and visual attention, they only explain the results in specific stimulus. We could not predict the performance of clients conveniently for real world images. We need a tool, in which any simulation experiment can be replicated exactly, to estimate the performance in any task. (3) In recent years, a variety of computational models have been proposed to describe the data obtained from experiments or explain the cognitive mechanisms underlying performance in a task. An existing model could be adaptive to simulate and explain the process of visual attention. We have proposed the model for the second problem.
The purpose of this study is (1) to clarify a proposed model related with response time, and (2) to calculate a standard profile based on this proposed model for estimating visual attention. In the experiment for (1), the same visual stimulus was tested by human subjects and the model. A linear relationship was given by the comparison of the results. In the experiment for (2), the attention state of subjects was affected by alcohol. For clinical purpose, a case report was described. The effectiveness of the proposed estimation was examined. The results suggested that the proposed model could estimate visual attention.
It concluded that the proposed model could be developed to an adequate estimation for visual attention. This will be conducive to quantitative analysis of visual attention for psychologists and clinical practitioners.
|
|---|
