点云交互式分割调研
1 Golovinskiy, A., & Funkhouser, T. A. (2009). Min-cut based segmentation of point clouds. ICCV Workshops, 39–46. http://doi.org/10.1109/ICCVW.2009.5457721
方法： knn建图，min-cut， 可以自动，也可以进行交互式分割。
defining a penalty function that encourages a smooth segmentation where the foreground is weakly connected to the background, and minimizing that function with a min-cut.
在一个室外场景数据集上，进行前景背景的分割。
2014_A NEW FRAMEWORK FOR INTERACTIVE SEGMENTATION OF POINT CLOUDS
Graph cut 方法
Random walking 方法
Establishing point cloud connectivity in this way avoids the problem of disconnected components encountered using KNN connectivity in unorganized point clouds.
First, user-selected nodes within the graph are assigned a classification label. These ‘seed points’ are designated as belonging to a given segment (defect or background). The task of the algorithm is then to determine which label should be assigned to each of the remaining unlabeled nodes
举的例子都很简单。
2009_Graph Cut Based Point-Cloud Segmentation for Polygonal Reconstruction
Graph cut 方法，一个文物的地形提取， 50万个点，种子点只需要不到50个， 同时只区分地形和非地形，二分类。
2018_GemSketch: Interactive Image-Guided Geometry Extraction from Point Clouds
图像引导的点云交互式分割。
先前的交互式点云分割方法，多关注于室外场景，物体与背景分割，室内场景，平面分割。 鲜有语义分割，不规则物体的语义分割。