A pseudomatched filter model applied to the transient system of human vision

A pseudomatched filter configuration is presented as a model for the transient system of human visual perception. This model resembles the models proposed by W. Reichardt (1971) and C. Rashbass (1970), but the pseudomatched filter model is linear. The choice of the model is based on the shape of the psychophysically measured impulse response and on physiological and functional arguments. The parameters of the model are estimated for different data sets. The model was fitted to experimentally determined impulse responses of the transient channel of the visual system. In most cases the model´s parameters were found, so a reasonable fit was obtained. It was found that under the same conditions (background level and field diameter) the estimated model was essentially the same for different subjects. Comparison of the results from the fits at different background levels showed that from the three parameters in the model only the angular frequency changed by more than the estimated standard deviation. It is shown that the model is in agreement with other psychophysical data. This is shown by predicting the threshold-versus-duration curve and the sensitivity to sinusoidal stimuli. The model´s threshold-versus-duration curves agree well with the experimental data.