An improved content-based image retrieval system based on a quantitative analytical hierarchical process with continuous learning

摘要

An improved content-based image retrieval system based on a quantitative analytical hierarchical process with continuous learning, including:a data collection unit to capture a given set of characteristics and at least one initial decision class with different values;a preprocessing unit for dividing the set of characteristics into a plurality of groups, evaluating initial priority values, dividing the set of characteristics in such a way that in each group the ratio of the number of values in each of the classes is inverse to the number of classes;a unit of calculation for evaluating weights of the analytical hierarchical process (AHP) for the majority of groups; anda central processing unit to form the final decision weights after incremental learning from the weights of each group, the central processing unit being configured to continuously monitor the size of the dataset and re-apply the analytical hierarchical process to the dataset; if the size of the record exceeds a certain threshold.