The Study Of Visual Effectiveness And Parameter Optimization Based On Gray

Effectiveness is a prerequisite for all design to serve us,Nowadays visual effectiveness is valued by more and more researchers. It is influenced by many Factors, When comes to a specific research process, Investigator usually choose a specific direction among the physical properties of visual stimuli, environmental observation and observer as an entry point. This study is based on gray-scale vision, Using the psychophysics experiments theory combine with modeling and optimization. Discuss the Cross-impact relationship between perspective, background gray, Simultaneous visual contract ratio and the visual effectiveness. constructing a visual effectiveness prediction model. establish the visual effectiveness model which can analyze the value data required by the user and output the physical attribute parameter date of the visual stimulation model. According to the analysis result of the visual effect factors, the thesis extracts the visual stimulus information, which is the key of research. Use the visual psychophysics and artificial intelligence research methods to architecture the framework of the gray visual effectiveness. It uses the improved limit method to predict different gray scale and uses different perspectives of visual to contrast threshold scale. As a consequence of threshold prediction data, the thesis uses constant stimulus method to take specific speculation. According to experimental data, it analyses the relationship of effectiveness among the visual angle, background gray scale, visual gray scale and visual validity.From the perspective of the artificial intelligence, the thesis uses the BP neural network to constructs the prediction model of visual effectiveness taking the gray scale value of the angle, background and object target as the network input while making the visual effectiveness as the network output. Mor eover, the research chooses another 10 different visual models which are selected to validate the experiment results. It verifies the generalization ability of the BP net.The BP neural network model will be optimized from the perspective of genetic algorithm, and its solution will be searched in the 3 dimensional visual physical properties parameter space. According to visual effectiveness value of users which they demand, the paper processes to fit and picks out the minimum error of visual value fitting effectiveness and the error to correspond with 3 dimension material physical property parameters which is as the model output. The parameters of 8 model groups which are as output are selected to verify the visual model of the experiment. And corresponding error is obtained, which is as to be the reference of the verification model. The results show that the optimized model has a certain fitting ability, and its output parameters of physical property can be used as reference and guidance for actual design.