Title :
Optimized Motion Energy Estimation for Group of Pictures in Multi-Level Error Protection of H.264/AVC Video Bitstreams
Author :
Huu Dung Pham ; Vafi, Sina
Author_Institution :
Sch. of Eng. & Inf. Technol., Charles Darwin Univ., Darwin, NT, Australia
Abstract :
A multi-level error protection technique for H.264/AVC video bitstreams structured by Group of Pictures (GOP) is proposed. Protection levels are determined by motion energy, which considers displacements of all macroblocks between consecutive frames. Lagrange polynomial is applied in predicting motion energy of a GOP, which is derived from previous GOPs having a high correlation with each other. The optimum number of previous GOPs is selected in order to obtain a good motion energy estimation of the current one. Conducted analyses and simulation results show that video performance obtained by the proposed technique outperforms other motion activity-based Unequal Error Protection (UEP) techniques, while a similar channel code rate is applied.
Keywords :
correlation methods; image motion analysis; polynomials; video coding; GOP; H.264/AVC video bitstreams; Lagrange polynomial; UEP techniques; channel code rate; correlation; group of pictures; motion activity-based unequal error protection; multilevel error protection technique; optimized motion energy estimation; protection levels; video performance; Delays; Error correction codes; Estimation; Polynomials; Vectors; Video coding; Group of Pictures (GOP); Lagrange polynomial; motion energy; unequal error protection;
Journal_Title :
Multimedia, IEEE Transactions on
DOI :
10.1109/TMM.2014.2348712