Creating 3D video content from existing 2D video has been stimulated by recent growth in 3DTV technologies. Depth cues from motion, focus, gradient, or texture shading are typically computed to create 3D world perception. More selective attention might be introduced using manual or automated methods for entertainment or educational purposes. In this paper, we propose an adaptive conversion framework that combines depth and visual saliency cues. A user study was designed and subjective quality scores on test videos were obtained using a tailored single stimulus continuous quality scale (SSCQS) method. The resulting mean opinion scores show that our method is favored by human observers in comparison to other state-of-the-art conversion methods.
展开▼
