45,99 €
inkl. MwSt.
Versandkostenfrei*
Versandfertig in 6-10 Tagen
payback
23 °P sammeln
  • Broschiertes Buch

This work presents an innovative approach to unequal protection for video transmission over noisy channels, based on a perceptual multi-criteria evaluation integrating three complementary dimensions of data importance.The temporal dimension classifies images according to their position in the GOP. The dynamic dimension identifies macroblocks containing significant motion, natural attractors of visual attention. The spatial dimension exploits computational models of attention to detect regions of visual interest.The originality lies in the Boolean superposition fusion methodology, which…mehr

Produktbeschreibung
This work presents an innovative approach to unequal protection for video transmission over noisy channels, based on a perceptual multi-criteria evaluation integrating three complementary dimensions of data importance.The temporal dimension classifies images according to their position in the GOP. The dynamic dimension identifies macroblocks containing significant motion, natural attractors of visual attention. The spatial dimension exploits computational models of attention to detect regions of visual interest.The originality lies in the Boolean superposition fusion methodology, which combines these three criteria according to logical rules adapted to the image's temporal class. This approach generates a global importance index guiding the differentiated allocation of error-correcting codes at macroblock granularity.Experimental validation shows gains of +2.43 dB in PSNR (+7.6%) compared with equal protection, with a controlled additional throughput cost (+2.9%). The approach stands out for its perceptual anchoring and methodological transparency, opening up prospects for adaptation to contemporary technologies.
Autorenporträt
Ouafae SERRAR (CRMEF Marrakech) i Oum El Kheir ABRA (CRMEF Rabat), eksperci w dziedzinie technologii edukacyjnych. Ich badania ¿¿cz¿ multimedialne systemy informacyjne i innowacje edukacyjne. Obecnie badaj¿ sztuczn¿ inteligencj¿ pod k¿tem adaptacji i personalizacji uczenia si¿.