摘要：为了准确地提取虚拟内窥镜的漫游路径,消除漫游过程中视觉上的画面旋转的现象,提出一种结合骨架提取与路径优化的虚拟内窥镜自动漫游方法.首先采用一种基于体素密度值加权的最短路径生成算法,在细化后的三维医学影像数据初始骨架上提取简单漫游路径;然后采用三次B样条曲线拟合方法得到最终的漫游路径;在沿此路径漫游时,通过计算最小旋转标架来优化虚拟摄像机的向上矢量,并设定向上矢量角度旋转阈值对漫游路径进行动态重采样,优化虚拟摄像机的位置参数.实验结果证明,该方法能提取准确的虚拟内窥镜漫游路径,并能获得较好的漫游视觉效果.

摘要：Parameterization is one of the key problems in the construction of a curve to interpolate a set of ordered points. We propose a new local parameterization method based on the curvature model in this paper. The new method determines the knots by mi- nimizing the maximum curvature of quadratic curve. When the knots by the new method are used to construct interpolation curve, the constructed curve have good precision. We also give some comparisons of the new method with existing methods, and our method can perform better in interpolation error, and the interpolated curve is more fairing.

摘要：Superpixels generation is becoming increasingly popular as a preprocessing in many computer vision applications. A superpixel is an image patch which has uniform pixels intensity and is aligned with intensity edges. Superpixels provide a convenient primitive from which local image features can be computed. So far, there are many methods to generate superpixels. Several main superpixels generation algorithms are summarized in this paper and the advantages and disadvantages of them are analyzed simply. In the end, some applications of superpixels are listed.

摘要：容积泵是一种通过改变密闭空间的容积来实现流体输送的泵。密闭空间容积的改变是通过围成它的机械部件之间的相对运动实现的。目前的容积泵无法避免这些机械部件之间的滑动摩擦。如果能设计出使用滚动摩擦代替所有滑动摩擦的结构,则可以让摩擦力大大减小,从而提高泵的效率和寿命。避免滑动摩擦的同时保证密封却是一个很有挑战性的几何设计问题。以正八面体为基本结构,通过在它的棱和面上放置特殊形状的转子和活塞,成功实现这种构造,并使用三维几何建模软件模拟了这种结构的工作过程。该工作不仅从理论上证实了无滑动摩擦的容积泵的可实现性,对高效率容积泵的设计也有着很强的应用价值。

摘要：提出一种隐式曲面多边形化的方法,将隐式曲面的多边形化分为2个阶段:首先根据法向约束对隐式曲面进行采样,得到稳定的采样粒子表示;然后在每个采样粒子处沿法线正负方向分别在隐式曲面内部和外部延伸一段距离,得到2个曲面法向附加点.将法向附加点和采样顶点进行四面体化,删除法向顶点及其相关联的边,最终得到隐式曲面的三角形网格模型.最后用实例表明了该方法的有效性.

摘要：针对现有口腔扫描设备获取的网格数据存在自相交、孔洞等一系列问题,提出一种快速、有效的自相交检测算法.为了降低全局网格自相交检测的时间复杂度,引入一种特殊的八叉树算法,利用细分的思想把三角面片放置在不同的立方体内以实现快速自相交检测,并通过删除操作消除错误连接;为了修复上一步骤中形成的大量网络孔洞,设计一种特殊的双边数据结构,能够快速有效地检测单连通封闭孔洞,并正确处理多个孔洞共用同一顶点的问题;孔洞修复完成后,利用拉普拉斯平滑策略对孔洞区域进行处理,获得过渡自然平滑的补洞效果.在80套口腔内扫描数据上的实验结果表明,所提算法不仅能够快速检测网格数据的自相交区域,而且能够高效完成补洞操作;与现有算法相比,在网格数量大于100万的数据上,处理速度提高大约10倍.

摘要：针对高分辨率的锥形束计算机断层扫描图像难以获取的问题,提出一种基于先验生成对抗网络的锥形束计算机断层扫描图像超分辨率方法,其通过微型断层扫描图像作为参照,对超分辨率网络进行弱监督训练.首先训练一个生成对抗网络,用于生成高质量的单颗牙齿的微型计算机断层扫描图像,并将其嵌入到一个U型的主干网络中作为先验解码器;然后利用低分辨率的多颗牙齿的锥形束计算机断层扫描图像对主干网络进行训练,先定位到锥形束计算机断层扫描影像中的每一颗牙,再分别提高分辨率;最后解决锥形束计算机断层扫描与微型计算机断层扫描之间的域差距,通过设计基于小波变换噪声提取的域适应退化模块,间接优化生成器生成更符合微型计算机断层扫描信息分布的图像.在锥形束计算机断层扫描数据集上进行实验的结果表明,与现有的超分辨率方法相比,所提方法的峰值信噪比提高了0.79~6.02 dB,感知相似度评价指标降低了0.01~0.72,并且在牙齿部分获得了更好的视觉效果,具有较强的竞争力.

摘要：人体动作识别因其难以结合时空域信息成为计算机视觉方向中一项具有挑战性的任务.提出一个多注意力时空图卷积网络,其核心思想是根据时间序列信息和人体骨架的自然连接构建一个连通图,然后利用具有多注意力机制的时空图卷积网络自动地学习空间和时间特征并且优化该连通图,最后实现对人体动作的预测.引入图注意力模块,模型构建的图的拓扑结构在初始化后会随着网络训练的过程进行优化,最终得到更适合表达人体动作的拓扑结构.此外,加入通道注意力模块,使网络能够更加注意相对重要的通道信息,从而提取更有效描述动作的特征.在公认的大型数据集NTU-RGBD和Kinetics上进行了大量的实验,结果表明该方法具有更高的识别准确率.

摘要：In this paper we present a new image zooming algorithm based on surface fitting with edge constraint. In surface fitting,we consider not only the relationship of corresponding pixels between the original image and the enlarged image, but also the neighbor pixels in the enlarged image according to the local structure of original image. Furthermore, during surface fitting, more interpolation constraints are used in the new algorithm for improving the precision of the super sampling pixels. The experimental results show that the new method outperforms the previous methods which based on surface fitting.

摘要：Triangulation is widely used in scientific research, such as finite element mesh generation, surface reconstruction and the reconstruction of the density field data. This paper proposes a new method combining image processing and density-controlled Centroidal Voronoi tessellations to quickly generate a density-controlled constrained Delaunay triangulation lbr 2D sea area. Firstly, preprocess digital images of the sea area and extract the boundary of seawater region by eight-neighbor searching algorithm. Then, 1 use Odd-Even Testing＂ method to check if one random vertex is inside the boundary and insert random vertices into the boundary. Finally, we get the CDT of random vertices by density-controlled CVT-Lloyd method. We also give some comparisons with existing methods, and our method performs better in final restllt of triangulation.