TEXTURE SIMILARITY METRICS APPLIED TO HEVC INTRA PREDICTION - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2015

TEXTURE SIMILARITY METRICS APPLIED TO HEVC INTRA PREDICTION

Résumé

Textures represent regions of homogeneous properties (e.g. color, directionality, edge distribution etc.). They cover wide areas of the visual scene with less importance as compared to structures. The typical image/video encoders aim at optimizing the bitrate within a certain distortion level. The distortion is usually measured via comparing pixel values, which could highly deviates from the perceived distortions especially for texture components. In this paper, we review and verify our recent work on using texture similarity metrics as a measure of perceived distortion in HEVC, and optimize encoder accordingly. Experimental results reveal the same findings in [1] [2] that when a texture similarity metric is used, the visual quality of the decoded textures is improved as well as the rate-similarity performance.
Fichier principal
Vignette du fichier
Naser_Ricordel_LeCallet_SIFWICT2015.pdf (8.86 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01164951 , version 1 (18-06-2015)

Identifiants

  • HAL Id : hal-01164951 , version 1

Citer

Karam Naser, Vincent Ricordel, Patrick Le Callet. TEXTURE SIMILARITY METRICS APPLIED TO HEVC INTRA PREDICTION. The third Sino-French Workshop on Information and Communication Technologies, SIFWICT 2015, Jun 2015, Nantes, France. ⟨hal-01164951⟩
591 Consultations
171 Téléchargements

Partager

Gmail Facebook X LinkedIn More