Abstract:The paper presents the concept of dynamic Contextual Augmented Reality Environments (CARE), in which augmentation presented to users is dynamically constructed based on four semantically described elements. The first element is the user's context (preferences, privileges, location, time, device's capabilities). The second element is a set of trackables — visual markers representing real world objects that can be augmented for a given user in a given context. The third element are content objects, representing interactive 2D and 3D multimedia content including video sequences and sounds to be presented on the trackables. The last one is a description of a user interface, which may be specific to a concrete device or application and which indicates the forms of information presentation and interaction available to a user.