Department of Electrical and Computer Engineering, Memorial University, St. John's, NL A1B 3X5, Canada.
Sensors (Basel). 2018 Jun 11;18(6):1900. doi: 10.3390/s18061900.
We present a method of segmenting human parts in depth images, when provided the image positions of the body parts. The goal is to facilitate per-pixel labelling of large datasets of human images, which are used for training and testing algorithms for pose estimation and automatic segmentation. A common technique in image segmentation is to represent an image as a two-dimensional grid graph, with one node for each pixel and edges between neighbouring pixels. We introduce a graph with distinct layers of nodes to model occlusion of the body by the arms. Once the graph is constructed, the annotated part positions are used as seeds for a standard interactive segmentation algorithm. Our method is evaluated on two public datasets containing depth images of humans from a frontal view. It produces a mean per-class accuracy of 93.55% on the first dataset, compared to 87.91% (random forest and graph cuts) and 90.31% (random forest and Markov random field). It also achieves a per-class accuracy of 90.60% on the second dataset. Future work can experiment with various methods for creating the graph layers to accurately model occlusion.
我们提出了一种在深度图像中分割人体部位的方法，当提供身体部位的图像位置时。目的是方便对大量人体图像数据集进行逐像素标注，这些数据集用于训练和测试姿势估计和自动分割算法。图像分割的一种常见技术是将图像表示为二维网格图，每个节点代表一个像素，相邻像素之间有边。我们引入了一个具有不同节点层的图来模拟手臂对身体的遮挡。一旦构建了图，带注释的部分位置就可以用作标准交互分割算法的种子。我们的方法在包含人体正面深度图像的两个公共数据集上进行了评估。与随机森林和图割（87.91%）和随机森林和马尔可夫随机场（90.31%）相比，它在第一个数据集上产生了 93.55%的平均每类准确率。它在第二个数据集上也实现了 90.60%的每类准确率。未来的工作可以尝试使用各种方法来创建图层，以准确地模拟遮挡。
