A HOG-SVM Based Fall Detection IoT System for Elderly Persons Using Deep Sensor

The population of elderly persons continues to grow at a high rate, and fall accidents in elderly persons have become a major public health problem. Highly developed IoT technology and machine learning enable the use of multimedia devices in a wide variety of elderly person’s protection areas. In this paper, a HOG-SVM based fall detection IoT system for elderly persons is proposed. To ensure privacy and in order to be robust to changes of the light intensity, deep sensor is employed instead of RGB camera to get the binary images of elderly persons. The persons are detected and tracked by Microsoft Kinect SDK, and the unwanted noise is reduced by noise reduction algorithm. After obtaining the denoised binary images, the features of persons are extracted by histogram of oriented gradient and the image classification is performed for judging the fall status by the liner support vector machine. If a fall is detected, the IoT system sends alert to the hospital or family members. This study builds a data set which includes 3500 images, and the experimental results show that the proposed method outperforms related works in terms of accuracy.