Image Mining

Dr.K.ThirugnanaSambanthan, Assistant Professor, SIMS

 Image mining is an important principle for gaining knowledge of images. It collects patterns of images which satisfy the properties of a given image and these patterns are useful to retrieve images.  The mining system incorporates computer vision, artificial intelligence and database. It finds all patterns without additional knowledge of image. It handles image notation in databases produced by indexing methods. It  is also a complicated process since it combines distinctive methods and the success of Image mining depends on the proficient approach yielding patterns and knowledge of images.

 Mining system considers problems geometric transformation, noise and other distortions of images. The images are obtained by semantic image interpretation and the results of mining can be checked with the model description and features which may be symbols or verbal terms.

 EXISTING IMAGE RETRIEVING SYSTEMS

Retrieval of image is through the description of images many image retrieval systems; say Google or Yahoo search images encompasses metadata.. It has to traverse a huge collection of digital images for identifying resultant image. Sometimes image similarity is motivated to archive images and also text based image filtering clinch image acquisition from web tags are defined with images to fulfill the venture of getting images. Many algorithms are making use of visual and text metadata for searching.    Image retrieval is also effectuated by Query By Image Content (QBIC). The problems of searching digital images in huge databases are resolved by this principle. This approach is a complement to traditional image processing system. Content Based Image Retrieval (CBIR) is another name of QBIC. Human cannot manually annotate the image by entering the keywords since data is very large and the resultant image is not the desired one.

 Image Mining software is the need of the hour as more searches are done using images and regarding image.