Visual pattern recognition : a preliminary investigation

No Thumbnail Available
Authors
Howard, Stephen O.
Advisor
Issue Date
1973
Keyword
Degree
Thesis (M.A.)
Department
Other Identifiers
Abstract

Several visual recognition models have been proposed which explain the results of a variety of perceptual and recognition tasks. Each task used reaction-time data as the response indicator. Recently, support for one of the models has come from a study analyzing the location of eye movement fixations. By using a serial experimental design, i.e., the object could not be seen in a single fixation, the experimentors found the presence of orderly, sequential fixation paths and presented a theory of a serial memory search, similar to one proposed and followed by researchers using reaction-times.While the existence of a serial process is not to be questioned, other conclusions which support the theory are. Recognition was only subjectively defined in the study, which begs the major question: are the eye movement researchers and reaction-time researchers studying the same process? Because no instructions were given to the subjects in the eye movement experiment, one wonders if eye movements would continue to be orderly and sequential when the solution to a discrimination problem is contingent upon recognition.Specifically, this experiment was designed to test the influence that the structure of experimental design had on a difficult recognition task, a modified replication of the eye movement design, and to test the prediction of the serial search model, supported by the eye movement data, using reaction time data.If the results from this research project support the predictions, they will be of further theoretical value by increasing the understanding of the complicated perceptual processes. If eye movements are an integral part of recognition, the theory will realize practical advantages, too. For instance, perceptual problems related to reading and emotional disturbances may be diagnosed by monitoring eye movements, as well as be eliminated by eye movement conditioning. If this research is not performed, the theory will not have received proper examination before its acceptance.

Collections