In this paper, a novel video coding technique is presented, based on the decomposition of the sequence spanned hyperspace in orthogonal components. Good quality video can be reconstructed at the receiver, particularly for very low bit rate applications, with low residual error and smooth appearance both in motion reconstruction and in the complete absence of crisp coding artifacts such as blocking structures, basing on a few proper selected examples of the speaker and a multidimensional interpolation algorithm. Principal component analysis exploited by singular value decomposition has been used to decompose the video sequence into eigen-images. A limited set of eigen-images is chosen to span the whole video sequence subspace. Only coefficients to be used in the interpolation procedure are transmitted along the communication channel. Preliminary results are presented for Akyio QCIF video sequence.
Constant bit rate Video Coding based on SVD video decomposition / Guaragnella, Cataldo; Politi, Tiziano. - (2007), pp. 113-116. (Intervento presentato al convegno IEEE International Conference on Virtual Environments, Human-Computer Interfaces, and Measurement Systems, VECIMS 2007 tenutosi a Ostuni, Italy nel June 25-27, 2007) [10.1109/VECIMS.2007.4373939].
Constant bit rate Video Coding based on SVD video decomposition
GUARAGNELLA, Cataldo;POLITI, Tiziano
2007-01-01
Abstract
In this paper, a novel video coding technique is presented, based on the decomposition of the sequence spanned hyperspace in orthogonal components. Good quality video can be reconstructed at the receiver, particularly for very low bit rate applications, with low residual error and smooth appearance both in motion reconstruction and in the complete absence of crisp coding artifacts such as blocking structures, basing on a few proper selected examples of the speaker and a multidimensional interpolation algorithm. Principal component analysis exploited by singular value decomposition has been used to decompose the video sequence into eigen-images. A limited set of eigen-images is chosen to span the whole video sequence subspace. Only coefficients to be used in the interpolation procedure are transmitted along the communication channel. Preliminary results are presented for Akyio QCIF video sequence.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
