Ecole Polytech. Fed. de Lausanne, Lausanne;
directed graphs; image motion analysis; image representation; image segmentation; statistical analysis; tree searching; affine motion models; bottom-to-top approach; breadth-first search strategy; directed acyclic graph; dynamic scenes; global depth ordering; local pair-wise relation; longest path problem; motion layer segmentation; pair-wise occlusion relations; pixel voting statistics; pseudo 3D representation; region segmentation; temporal continuity; temporally consistent layer depth ordering; 2½D representation;
机译:用于3D视频生成和编码的时间一致性深度图估计
机译:基于深度图像表示的具有触觉交互的3D视频播放器系统
机译:基于深度卷积神经网络和时空条件随机场的时间一致深度图预测
机译:通过像素投票用于伪3D表示的时间一致的层深度排序
机译:具有高时间,横向和深度分辨率的计算3D成像
机译:使用双层混合超颖表面产生高色彩深度的超紧凑型超像素
机译:使用分层深度图像和修复的时空一致的基于深度图像的渲染