Neural network model of the binocular fusion in the human vision

This paper proposes a model of the binocular fusion based on the psychological experimental results and the physiological knowledge. Considering the psychological results and the physiological structure, we assume that the binocular information are processed by several binocular channel having different spatial characteristics from low spatial frequency to high spatial frequency. In order to examine the mechanism of the binocular fusion, we construct a five layer neural network model, and train it by the back-propagation learning algorithm with use of psychological experimental data. After completion of learning, the generalization capability of the network are examined. Further, the response functions of the hidden units have been examined, which suggested that the hidden units have spatial selective characteristic. In other words, the binocular information is considered to be pressed of parallel channels with different spatial characteristics.