A Novel Image Structural Semantics Modeling

In this paper, we propose a novel image structural semantics model. The detailed process of mathematical inference for the whole system is developed. The model applies the knowledge of topology, neural networks, and psychology. The system first builds a structure space T using the structure parameters extracted from an image or video segment. Then T is mapped to a structure information space I_T by a transformation function set phi which is based on psychometric functions. The model introduces a new regulation based on psychology and neural networks for semantic analysis of image or video databases, and we call it psychological response neural network