他使用 GPU 遮挡查询获取可见性信息，利用 HLOD，快速更新层次结

他使用 GPU 遮挡查询获取可见性信息，利用 HLOD，快速更新层次结构中节点的可见性信息和细节模型[7]。Havran.V等对于kd-tree构建过程中的终结条件进行了研究，提出了一种自动终结算法的策略，同时还使用了一种新的分割剪裁算法，在牺牲了少许构建速度的情况下，大幅度提高了kd-tree的质量[8]。Wald I等在研究和比较了几种kd-tree构建算法后，提出了一种新的算法，将场景中的队形边界按照分割轴的坐标值进行排序，继而将传统的SAH算法复杂度由O(nlog2n)降低到了O(nlogn)[9]。Jiri Bittner 等对时间连续性和空间连续性进行了深入的研究，结合空间划分技术，提出了一个高效的基于图形硬件的遮挡剔除算法[10]。

結果 (英文) 1:

His visibility using GPU occlusion queries to obtain information using HLOD, quickly update the visibility of nodes in the hierarchy of information and details of models [7]. Havran.V for KD-tree in the construction of the final conditions were studied, puts an automatic end strategy, while also using a new clipping algorithm of segmentation, in the case of sacrificing a little building, substantially improves the quality of KD-tree [8]. Wald I built research and comparison of several KD-tree algorithm, a new algorithm is presented, formation boundaries in accordance with split axis coordinate values are sorted, then the traditional SAH complexity o (nlog2n) down to o (nlogn) [9]. Jiri Bittner on continuity and in-depth study of spatial continuity, combined with space technology, this paper presents an efficient graphics hardware occlusion culling algorithm [10].

結果 (英文) 2:

He uses GPU occlusion queries to obtain information visibility, use HLOD, visibility quickly update information and details model hierarchy of nodes [7]. Havran.V etc. For kd-tree building process termination conditions studied, we propose a strategy for the automatic termination algorithm, but also uses a new algorithm for segmentation cut, at the expense of a little under the circumstances to build speed substantially increased the kd-tree quality [8]. Wald I, etc. After studying and comparing several kd-tree construction algorithm, we propose a new algorithm, the scene boundary formation by dividing axis coordinate values ​​are sorted, then the traditional algorithm complexity by SAH O (nlog2n) down to O (nlogn) [9]. Jiri Bittner and other right temporal continuity and spatial continuity conducted in-depth research, combined with space partitioning technique, proposed an efficient graphics hardware-based occlusion culling algorithm [10].

結果 (英文) 3:

He used GPU occlusion query to obtain visibility information, using HLOD, quick update node in the hierarchy of visibility information and details of model [7]. Havran.V were studied for kd-tree in the process of building the termination condition, put forward a kind of automatic termination algorithm strategy, but also the use of a new segment clipping algorithm, at the expense of some construction speed circumstances, greatly improves the quality of [8] kd-tree. Wald I in the study and comparison of several kd-tree construction algorithm, a new algorithm is proposed, the boundary in the scene according to the coordinate axis formation segmentation value sorting,Then the traditional SAH algorithm complexity from O (nlog2n) is reduced to O (nlogn) [9]. Jiri Bittner on the time and spatial continuity of in-depth study, combined with space partition technique, propose a graphics hardware occlusion culling algorithm based on [10] efficiency.