signed

QiShunwang

“诚信为本、客户至上”

Fall detection for elderly people using the variation of key points of human skeleton

2021/4/26 21:06:52   来源:

ABSTRACT In the area of health care, fall is a dangerous problem for aged persons. Sometimes, they are a serious cause of death. On the other hand, the number of aged persons will increase in the future. Therefore,it is necessary to develop an accurate system to detect fall. In this paper, we present spatiotemporal method todetect fall form videos filmed by surveillance cameras. Firstly, we computed key points of human skeleton.We calculated distances and angles between key points of each two pair sequences frames. After that,we applied Principal Component Analysis (PCA) to unify the dimension of features. Finally, we utilizedSupport Vector Machine (SVM), Decision Tree, Random Forest and K Nearest Neighbors (KNN) to classify features. We found that SVM is the best classifier to our method. The results of our algorithm are as follow: accuracy is 98.5%, sensitivity is 97% and the specificity is 100%.

在医疗保健领域,跌倒是老年人面临的一个危险问题。有时,它们是一个严重的死亡原因。另一方面,老年人的数量在未来将会增加。因此,有必要开发一套准确的跌倒检测系统。在本文中,我们提出了一种时空方法来检测监控摄像机拍摄的跌倒视频。首先,计算了人体骨架的关键点。我们计算了每一对序列帧的关键点之间的距离和角度。然后,我们使用主成分分析(PCA)来统一特征的维数。最后,我们利用支持向量机(SVM)、决策树、随机森林和K近邻(KNN)对特征进行分类。我们发现支持向量机是我们方法的最佳分类器。结果表明:该方法的准确率为98.5%,灵敏度为97%,特异性为100%。

In this section, we present our proposed method to detect fall. Firstly, we extracted features from the videos. Indeed, extracting features is devised to two steps: we detect key points of the human body’s skeleton. We used a 2D skeleton model to detect key points. Our challenge is using just simple RGB camera to detect fall. After that, we use these key points to compute the change of distance and angle between the same key points into each two pair sequential frames. After that, we apply Principal Component Analysis (PCA) to unify the size of the videos. Finally, we classify the features that we have computed to detect fall in the video.

作者再文中提出的检测摔倒的方法。首先,我们从视频中提取特征。事实上,提取特征被设计成两个步骤:我们检测人体骨骼的关键点。我们使用二维骨架模型来检测关键点。我们的挑战是使用简单的RGB相机来检测摔倒。然后,我们利用这些关键点来计算相同关键点在每一对序列帧中的距离和角度的变化。然后,我们使用主成分分析(PCA)来统一视频的大小。最后,我们对我们计算的特征进行分类,以检测视频中的跌倒。