基于感知的室内场景的多视点渲染优化

研究了室内场景的自适应多视点渲染优化问题. 对于输入的三维室内场景模型，首先定义了一个新颖的基于几何信息、结构信息及美学信息的视点感知度量函数; 进而建立了一个多视点整体优化模型; 然后利用模拟退火算法优化得到用户指定个数的多视点集合，最后渲染出图. 实验结果表明，对于室内场景多视点渲染优化这一问题，提出的方法能够自适应地得到一系列全面且美观的渲染图片.
Abstract：Multi-view rendering optimization for indoor scenes in the process of interior design was studied. Given an indoor scene model，firstly，a novel and perception-aware viewpoint metric function that is based on geometry information，structure information and aesthetic information was defined. Secondly，a multi-view objective function was modelled and optimized to obtain a multi-view set by using the simulated annealing algorithm. Extensive experiments show that the proposed method can adaptively obtain a set of all-sided and artistic rendering images for multi-view rendering optimization.