Semantic-Based Web Mining For Image Retrieval Using Enhanced Support Vector Machine

This paper deals with the semantic-based web mining for image retrieval by means of enhanced Support Vector Machine (SVM). Generally, conventional Content-Based Image Retrieval (CBIR) systems are unsuccessful to satisfy users’ requirement because of the ‘semantic gap’ among the derived features and the user’s query. A large amount of existing approaches shows certain predetermined semantic category and allocate the images to suitable categories through certain learning processes. In contrast, these approaches constantly require human involvement and depend on content-based features. Here, semantic-based web mining for image retrieval by means of enhanced Support Vector Machine (SVM) is introduced. The outcome of this text mining process includes two maps which expose the semantic associations among images and keywords, respectively. These maps are employed to carry out image retrieval processes. The experimental results reveal the effectiveness of the enhanced SVM and high retrieval accuracy and relevance of retrieved images.