One of the major challenges in content-based multimedia retrieval is due to the problem of semantic gap encountered as a consequence of significant disparities between inherent representational characteristics of multimedia and its meaningful content sought by the user. This work concerns the advancement of the semantic augmentation techniques focused on bringing together low-level visual representation of multimedia and its semantics thus attempting to alleviate the above semantic gap problem by augmenting the information used by a multimedia database system in order to improve the efficiency of access and retrieval. The main emphasis and contributions of this work summarized below are in the domain of supervised discriminative learning methods and ensemble techniques...