Abstract. In this paper a novel approach to metadata representation for semantic-based multimedia security and management is proposed. This approach is based on semantically meaningful object segmentation and digital data hiding to facilitate image indexing as well as identification for dedicated image databases. An image is considered as a collection of the regions that correspond to objects, where these objects are associated with some hidden metadata describing some of their features. The metadata is hidden inside the raw images themselves for security and "portability" reasons. The embedding of hidden metadata allows moving an image from one database to another, as well as the insertion/cropping of objects from one image to another, while still preserving the associated descriptions. The perceptual invisibility of the integrated metadata confirms the high performance of this proposed object-based hidden metadata representation.