Classifying Human Actions Using an Incomplete Real-Time Pose Skeleton
Classifying Human Actions Using an Incomplete Real-Time Pose Skeleton
复制标题
使用不完整的实时姿势骨架对人类动作进行分类
DOI：
10.1007/978-3-540-28633-2_119
复制
发表时间：
2004
期刊：
影响因子：
--
通讯作者：
G. West
中科院分区：
文献类型：
--
作者：
Patrick Peursum;H. Bui;S. Venkatesh;G. West
目前，大多数人类动作识别系统都是用没有缺失数据的特征集来训练的。不幸的是，使用人体姿态估计模型来提供更多描述性特征也需要增加对遮挡的敏感性，这意味着不完整的特征信息对于现实场景将是不可避免的。为了解决这个问题，我们的方法是将处理遮挡姿态数据的责任从姿态估计器转移到动作分类器上。这允许使用一个简单的、实时的姿势估计（简笔画），而不是估计它不能快速找到的肢体的位置。该系统通过背景减法来跟踪人们，并从他们的剪影中提取（可能不完整的）姿势骨架。然后使用隐马尔可夫模型来处理缺失数据，利用不完整的姿势特征成功地对几个人类行为进行分类。
Currently, most human action recognition systems are trained with feature sets that have no missing data. Unfortunately, the use of human pose estimation models to provide more descriptive features also entails an increased sensitivity to occlusions, meaning that incomplete feature information will be unavoidable for realistic scenarios. To address this, our approach is to shift the responsibility for dealing with occluded pose data away from the pose estimator and onto the action classifier. This allows the use of a simple, real-time pose estimation (stick-figure) that does not estimate the positions of limbs it cannot find quickly. The system tracks people via background subtraction and extracts the (possibly incomplete) pose skeleton from their silhouette. Hidden Markov Models modified to handle missing data are then used to successfully classify several human actions using the incomplete pose features.