MORSA: an algorithm to discover association rules in image data using recurrent items and significant rare items

Based on preliminary image processing, and content-based image retrieval technology, this paper presents MORSA, an algorithm for discovering association rules among image data using recurrent items and rare items in images stored in image data repository. The rare items to be considered are items that occur infrequently but that take place concurrently with some specific items that happen with frequency above a certain threshold.