MPEG-7 Description for Scalable Video Reconstruction

We present an MPEG--7 compliant description of video sequences for scalable transmission and reconstruction. The proposed method is content-based and permits efficient and flexible video coding while keeping the benefits of textual descriptions in database applications. Video objects are described in terms of shape, color, texture and motion. These features are extracted automatically and are sufficient in a wide range of applications like smart video cameras and Universal Multimedia Access (UMA). Scalable sequence reconstruction is obtained by providing at least one quantitative as well as a qualitative descriptor for each feature; video objects are described by a subset of the proposed descriptors. The scalability and the compactness of the description are demonstrated in an outdoor application.

Related material