Multimedia data mining

Data mining is a misnomer which means extracting knowledge from large amounts of data. The main steps in knowledge recovery are data cleaning, data integration, data selection, data transformation, data mining, pattern evaluation and knowledge presentation. In multimedia data mining, multimedia data base system stores and manages a large number of multimedia data which includes audio, video, image, graphics, speech, text, document etc. Popularity of this system in this current era is due to the sophisticated gadgets like digital camera, CD-ROMs and the internet. Consideration of two indexing and retrieval systems required in similarity searching in multimedia. They are description-based retrieval systems and content-based retrieval systems. Various approaches proposed in multimedia data mining are colour histogram-based signature, multifeature composed signature, wavelet-based signature and wavelet-based signature with region based granularity.