基于兴趣点统计特征的双人交互行为预测算法-《计算机技术与发展》

文章信息/Info

Title:: Human Interaction Prediction Algorithm Based on Statistical Features of Interest Points

Author(s):: JI Xiao-fei; XIE Xuan; School of Automation,Shenyang Aerospace University,Shenyang 110136,China

Keywords:: interaction prediction; statistical features of interest points; bag of words; Gaussian models; probability prediction

摘要:: 针对一些复杂敏感场景下需要快速及时地对人类交互行为做出预测的问题,提出了一种基于兴趣点统计特征的双人交互行为预测方法。该方法首先对动作视频提取时空兴趣点,并对其进行 3D-SIFT 描述,然后利用词袋方法对动作视频进行表示在训练阶段,利用高斯模型建立不同时间比例下每个动作的预测模型。在动作预测阶段,对于一个未知长度的动作视频,提取其词袋表示,并将其与所建立的不同时间长度的预测模型进行比较,得到与各模型之间的预测相似概率,最终实现对该交互行为的识别预测。利用 UT-interaction 数据库对该方法进行测试的实验结果表明,该方法易于实现,实时性好,并具有较好的预测效果。

Abstract:: A human interaction prediction algorithm based on statistical features of interest points method is proposed to solve the problem that human interaction needs to be predicted quickly and timely in some complex and sensitive scenarios. First,the spatio-temporal interest points are extracted and performed 3D-SIFT description,then the bag of words is used to represent the action video. In the training,Gaussian models are used to establish the action model for each action at different time scales. In the prediction,the bag of words representation is extracted and compared with the established prediction models of different time lengths to obtain similar prediction probabilities between the models for an action video of unknown length. Finally,the recognition and prediction of interaction prediction is completed. The experiment on UT-interaction dataset demonstrates that the proposed approach is easy to implement with better real-time performance and predictive effect.