Efficient Implementation of CBIR System and Framework of Fuzzy Semantics

Due to the enormous increase in image database sizes, as well as its vast deployment in various applications, the need for an efficient image retrieval system development arises. Present retrieval systems lack in some aspect or other to get accurate results. To retrieve the results on a subjective basis we have integrated all the feature vectors (namely color, texture and shape) and user can choose their weights relatively. To make the retrieval system more accurate and user interactive, relevance feedback is employed which improves the results drastically. Thus, the system becomes efficient by weighted integration of low level features and Relevance Feedback. Practical implementations of all the above techniques are done using MATLAB. Furthermore, one of the big challenges faced by content-based image retrieval (CBIR) is the 'semantic gap' between the visual features and the richness of human semantics for image content. We put forward a neural network approach to extract the image fuzzy semantics ground on linguistic expression based image description framework (LEBID). We have provided a framework using feed forward neural network to model the vagueness of human visual perception.