Effective feature extraction and image indexing for web image mining

Abstract

A good web image mining system is expected to provide users, the underlying knowledge and patterns in images which include image storage, image processing, feature extraction, image indexing and retrieval, pattern and knowledge discovery. So, Image mining systems needs an efficient mechanism for image retrieval conventional databases using Primary and Secondary indexes used currently the retrieval system for similarity hidden phase. In this case indexing has been done in similarity space. Indexing techniques ranging from standard methods such as inverted access methods and multidimensional methods. A specific indexing scheme is necessary for web image mining which presents and efficient color indexing scheme for similarity based retrieval system with the advancement in new image retrieval mechanism, gives better performance in retrieval techniques towards accuracy, time and memory utilization. Good indexing results in better classification and clustering of large image databases. Hence, the indexing mechanism is very important step in web image mining. Accessing newlinelarge image databases which are available in the web portal needs opted Indexing structure instead of reducing the contents of different kind of database to process quickly. It paves a path towards the increase in the number of efficient image retrieval techniques and numerous researches in the field of image indexing in large image data sets. Normally the image retrieval is facing difficulties like a) Merging the diverse representations of Images and its Indexing is a challenging one b) The low level visual characters and semantic characters associated with an image are indirect proportional one and c) Noisy and less accuracy in extraction of image information (semantic and predicted attributes). This research work clearly focuses and takes the base of reverse engineering and de-normalizing concept by looking back on, how the data can be stored effectively. So the retrieval becomes easy and too fast. The following works have been carried out and results