摘要 |
A system and method for categorizing non-textual subject data (14), such as digital images, utilizes content-based data and meta-data (16) to determine outcomes of classification tasks (54, 62, 64, 66, 68 and 70). The classification system (10) has a modular architecture in which modules (42) configured to perform specific functions, including algorithmic functions, can be integrated or deleted from the system (10). At the center of the classification system (10) is a decision module (30) comprising: (1) a task component (44) having a number of classification tasks (54, 62, 64, 66, 68 and 70) arranged within a task tree configuration (52 and 110) , (2) an algorithmic component (46) for selecting an algorithm for each classification task (54, 62, 64, 66, 68 and 70), (3) a sub-algorithmic component (48) for selecting sub-algorithmic routines (78, 80 and 82) for each algorithm, and (4) a learning component (50) for constructing and modifying the arrangement of the task tree (52 and 110) and the classification tasks (54, 62, 64, 66, 68 and 70) based on the frequencies of occurrences for the classes associated with a set of files (20). |