Stereoscopic 3D image quality assessment based on cyclopean view and depth map

Sid-Ahmed Fezza 1 Mohamed-Chaker Larabi 2
2 XLIM-SIC - SIC
Université de Poitiers, XLIM - XLIM
Abstract : This paper presents a full reference quality assessment metric for stereoscopic images based on perceptual binocular characteristics. To ensure that the predicted 3D quality of experience is as reliable and close as possible to 3D human perception, the proposed stereoscopic image quality assessment (SIQA) method is relying on the cyclopean image. Our approach is motivated by the fact that in case of asymmetric quality, 3D perception mechanisms place more emphasis on the view providing the most important and contrasted information. We integrated this psychophysical findings in the proposed 3D-IQA framework thanks to a weighting factor based on local information content. Add to that, to take into account the disparity/depth masking effect, we modulate the obtained quality score of each pixel of the cyclopean image according to its location in the scene. Experimental results show that the proposed metric correlates better with human judgement than the state-of-the-art metrics.
Type de document :
Communication dans un congrès
IEEE Fourth International Conference on Consumer Electronics - Berlin (ICCE-Berlin), Sep 2014, Liège, Belgium. pp.335 - 339, 2014, 〈10.1109/ICCE-Berlin.2014.7034289〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01159151
Contributeur : Mohamed-Chaker Larabi <>
Soumis le : mardi 2 juin 2015 - 16:34:46
Dernière modification le : jeudi 21 janvier 2016 - 10:00:29

Identifiants

Collections

Citation

Sid-Ahmed Fezza, Mohamed-Chaker Larabi. Stereoscopic 3D image quality assessment based on cyclopean view and depth map. IEEE Fourth International Conference on Consumer Electronics - Berlin (ICCE-Berlin), Sep 2014, Liège, Belgium. pp.335 - 339, 2014, 〈10.1109/ICCE-Berlin.2014.7034289〉. 〈hal-01159151〉

Partager

Métriques

Consultations de la notice

73