A novel arbitrary-block-shaped motion estimation technique is presented, able to deal with illumination variation in video sequences. The proposed technique allows adapting the search for the motion and illumination variation model on coherently moving pixels experiencing the same kind of illumination variation field. Experiments carried out on both synthetic and real video sequences show that the proposed technique allows us to significantly improve the accuracy of the estimation.
Improving Block–Based Motion Estimation in the presence of illumination variation / Guaragnella, Cataldo; DI SCIASCIO, Eugenio. - 2:(2002), pp. 555-558. (Intervento presentato al convegno 14th IEEE International Conference on Digital Signal Processing, DSP2002 tenutosi a Santorini, Greece nel July,1-3, 2002) [10.1109/ICDSP.2002.1028150].
Improving Block–Based Motion Estimation in the presence of illumination variation
GUARAGNELLA, Cataldo;DI SCIASCIO, Eugenio
2002-01-01
Abstract
A novel arbitrary-block-shaped motion estimation technique is presented, able to deal with illumination variation in video sequences. The proposed technique allows adapting the search for the motion and illumination variation model on coherently moving pixels experiencing the same kind of illumination variation field. Experiments carried out on both synthetic and real video sequences show that the proposed technique allows us to significantly improve the accuracy of the estimation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.