In the paper, a method for extracting interest points in images which is based on weighted local entropy is presented.
文中提出了一种基于加权局部熵的图像兴趣点检测方法。
Combined with weighted filtration to successful sample points, it uses the method of sampling in three-dimensional space and range constraint to acquire three-dimensional coordinates of nodes.
为了减小定位误差和提高算法的适应性，利用三维空间抽样和范围约束的方法，并结合对成功样本点的加权筛选，获得节点的三维估计坐标以实现抽样定位。
The new algorithm adds a direction vector to every depth pixel in LDI , and creates images from new vantage points by adopting a weighted rendering method for pixels with the same depth.
该算法对LDI的每个深度像素增加一个方向向量，对位于同一深度的像素采用加权平均的方法生成新视点下的目标图像。
应用推荐