Abstract: In this work, we propose a technique for facial expression recognition to bridge the semantic gap among the features that can be extracted in a contentbased video retrieval system. The paper aims to provide accurate and ...