A Real Time Object Detection System for Infant Safe Sleep Based on YOLOv5 Algorithm

Abstract

In this paper, a real-time object detection system is proposed to reduce the risk of Sudden Infant Death Syndrome (SIDS) related to unsafe sleeping positions. The main purpose is to apply the deep learning technology and train the YOLOv5 object detection algorithm to detect and recognize whether the safest sleeping positions. Experimental results show that the proposed model is able to achieve an accuracy of more than 99%, and the inference speed has reached 2.2 ms, which makes it compatible with real-time requirements. It can be integrated into baby monitoring devices, infant safety sleep detection systems, and mobile applications.