Stacked generalisation: A novel solution to bridge the semantic gap for content-based image retrieval

A two-stage mapping model (TSMM), which can be thought of as a two-levels stacked generalisation scheme for image classification, is presented. The model is proposed to bridge the semantic gap between low-level image features and high-level concepts in a divide-and-conquer manner, and aimed at minimising the gap by reducing classification errors. The idea is to design two level-0 generalisers to classify colour and texture features into colour and texture concepts respectively. Then, a level-1 generaliser is designed to classify the colour and texture concepts as middle-(words)-level concepts into high-level conceptual classes.