Este seminario está cofinanciado por el Proyecto PRO-MULTIDIS-CM (Programa S - 0505 / TIC / 0233, IV PRICIT, CM) y por el Programa Oficial de Posgrado en Ingeniería Informática y de Telecomunicación del Departamento de Ingeniería Informática de la Universidad Autónoma de Madrid.
Video segmentation is an integral part of many multimedia applications, such as query and retrieval from video databases, as well as an important step towards semantic video analysis and understanding. Although several approaches have been proposed in the literature for video segmentation, most of them operate in the uncompressed pixel domain. This provides them with the potential to estimate object boundaries with pixel accuracy but requires that the processed sequence be fully decoded before segmentation can be performed. It also leads to high computational complexity, as a result of the large number of pixels that have to be processed and the need for estimating motion features, that often arises. As a result, the usefulness of such approaches is usually restricted to nonreal-time applications, unless restrictive assumptions regarding camera motion or background homogeneity are made. To counter these drawbacks of pixel-domain approaches, compressed-domain methods have been proposed for spatiotemporal segmentation. These typically use directly the DCT coefficients of the compressed video, eliminating the need for compelete decoding. They also make use of motion information readily available in the compressed stream.
In this talk, the spatiotemporal segmentation of compressed video sequences will be discussed. Starting with a short discussion on the differences between pixel-domain and compressed domain segmentation, the talk will soon proceed to the presentation of the state-of-the-art in compressed video segmentation. Some methods of the recent literature will be discussed in detail and indicative results will be shown, to provide an indication of the performance of comressed domain methods and how this relates to the performance of pixel-domain ones. The talk will conlude with a discussion on current and future trends in compressed video segmentation, including the introduction of knowledge in the segmentation process.
Vasileios Mezaris received the Diploma degree and PhD in Electrical and Computer Engineering from the Aristotle University of Thessaloniki, Thessaloniki, Greece, in 2001 and 2005, respectively. He is a postdoctoral research fellow with the Informatics and Telematics Institute/Centre for Research and Technology Hellas, Thessaloniki, Greece. His research interests include image and video analysis, content-based and semantic image and video retrieval, ontologies, multimedia standards, knowledge-assisted multimedia analysis, knowledge extraction from multimedia, medical image analysis. He is a member of the IEEE and the Technical Chamber of Greece.