Stimulus correlates of visual pattern recognition: a probability approach.

Publisher Website
Google Scholar
Abstract
The first part of this paper consists of a discussion of the rules for generating visual stimulus material, called metric figures. Statistical properties are discussed with reference to general rules for generating figures, special sampling procedures, and the addition of two types of noise. Physical properties, e.g., size and orientation, size of samples, and associative values of metric figures, are dealt with separately. In the second part two experiments are reported in which some of the above-mentioned properties are examined in a recognition task. In the first experiment rules of sampling are shown to have a major effect on recognition, with differences due to orientation and type of figure depending on these rules. In the second experiment, in which three types of figures generated by sampling without replacement are presented in the four cardinal orientations, and after extra practice, mirror-imaged figures and vertical orientation of the figures'' base line are shown to be superior to other types of figures and to horizontal orientation. Method and findings are discussed with special reference to rules of sampling.

This publication has 3 references indexed in Scilit: