Studied the question of model orientation, cell block structure, sample combination, data search etc, and presented a few methods of sample combination and data search.
研究了模型定位及模型单元块构造、样品组合、数据搜索等问题，并且给出了样品组合及数据搜索的一些方法。
In order to accelerate detection speed, video images are divided into blocks. Each block motion orientation is estimated by block matching methods.
为了加快检测速度，将视频图像分割成大小相等的块，并估计每个块的运动方向。
In order to overcome the non-uniform phenomenon and the dependency on orientation when rendering, a block rule by marking the region of interest in objective space is presented.
为解决面标记策略中着色与切割时存在的着色不均和方向依赖性问题，本文通过在物体空间中标记感兴趣区域，提出块标记策略的实时着色与实时切割技术。
And a time sequence of motion orientation for each block is generated over a sliding time window. Then accumulation and main motion orientation are computed according to the sequence.
采用滑动时间窗口生成块运动方向时间序列，在此时间序列的基础上计算块的累积量和主运动方向。
And a time sequence of motion orientation for each block is generated over a sliding time window. Then accumulation and main motion orientation are computed according to the sequence.
采用滑动时间窗口生成块运动方向时间序列，在此时间序列的基础上计算块的累积量和主运动方向。
应用推荐