ID 原文 译文
38816 这一烟雾仿真算法基于格子玻尔兹曼方法实现，对动量分布函数，涡量分布函数以及温度分布函数进行离散化，通过三个分布函数和涡量方向场的迭代更新实现烟雾仿真。 Our smoke simulation model was based on Lattice Boltzmann method, which discretisizes momentum distribution function, vorticity distribution function and temperature distribution function. Smoke simulation was realized through updating of three distribution functions and vorticity direction field.
38817 与格子玻尔兹曼方法相同，该烟雾仿真算法通过分布函数项在格点之间的迁移仿真流体的流动过程，通过添加碰撞项和外力项仿真流体粒子之间的相互作用以及流体粒子受外力作用。 Our method simulated flow process through distribution function items moving among grid points and simulated interaction between fluid particles and external force added to particles with collision terms and external force terms separately in accordance with Lattice Boltzmann method.
38818 烟雾仿真实验结果证明涡量修正格子玻尔兹曼算法可以高效的生成高真实度的仿真图像。 Results of smoke simulation experiments prove that our method could simulate smoke motion with high realism and performance.
38819 该算法在单GPU加速的条件下在分辨率为64×64×128的网格上可以达到实时仿真的仿真效率。 As for grid resolution of 64×64×128, our algorithm could achieve real-time simulation with single GPU acceleration.
38820 体育视频包含大量不同类型的人体，其中运动员的行为与比赛进程和视频内容直接相关，因此运动员检测是体育视频分析的关键环节。 Sports video contains many different human bodies, in which players' behavior is directly related to the game and sport contents, therefore, accurate player detection is the key issues of sports video analysis.
38821 现有人体目标检测算法在通用人体检测任务上取得了良好的性能，但是无法有效区分运动员和非运动员。 Existing human detection algorithms have achieved good performance on general human detection tasks. However, these methods will detect player and non-player in sports videos at the same time and cannot further distinguish players from other human bodies.
38822 专门训练一个运动员检测模型需要标注大量的运动员位置，成本较高。 It is possible to train a specific model for player detection. However, it needs large amount of bounding box labels for players.
38823 本文提出了一种基于多示例学习的人体目标检测方法。 This paper proposes an object detection method based on multiple instance learning.
38824 在通用人体检测的基础上，引入多示例学习模块，基于图像级标注，通过弱监督方式自动学习获取特征映射矩阵，将人体特征映射到运动员特征空间，最后通过度量人体特征与运动员特征之间的相似度，实现运动员与非运动员的区分。 After the object detection for all kinds of human body, a multiple instance learning model is designed. It automatically obtains a feature mapping matrix in the way of weakly supervised learning from the training set with image-level labels. Finally, the player and non-player could be distinguished based on the similarity in the mapping feature space.
38825 对比实验结果表明，本文方法充分利用通用人体检测框架，以极小的标注数据量达到了专门训练运动员检测模型的精度。 The compared experimental results demonstrate that the proposed could utilized the existed human body detection framework effectively and detect the player with the minimum labeling cost. And the accuracy is same as the model specifically trained with the player bounding boxes.